talk-data.com
People (11 results)
See all 11 →
Activities & events
| Title & Speakers | Event |
|---|---|
|
AI Builders London Old St. - Mar 5
2026-03-05 · 17:30
🎟️ Get tickets: https://lu.ma/ai-builders 🎟️ ☝️This is a free meetup however a Luma ticket is required! Join our meetup for AI engineers & founders! We share the latest insights about: AI dev tools, Agent frameworks, RAG pipelines, automation hacks, Cursor/Claude code hacks, and new gen-AI models. **:: FOR WHO ::** ✅ Anyone actively building with Generative AI ✅ Devs, Product peeps, Data lovers, ML engineers, Founders ⚠️Technical LLM knowledge required!* :: FORMAT ::
**:: AGENDA ::** 17:30 🤝 Walk-in 18:00 🍕 Pizza (be early!) 18:30 🎤 Pioneer Speaker 19:00 ---- 💬 BREAK TIME ----- 19:30 💻 Demo (TBA) 19.45 💻 Demo (TBA) 20.00 🍻 Drinks 21.00 End :: FAQ :: • What is AI Builders? A self-organizing nonprofit community of 3000+ AI nerds🤓. and yes.. we're building a democratic AI CEO and run on opencollective.com donations. • Can I demo, give a talk, or just help out? Message Arthur (+31636570260) in case you want to shine✨ on stage and grow your network! • *I'm not technical. Can I come? Yes! To fully enjoy the meetup, we recommend chatting with AI to understand these basic LLM concepts: Multimodal, Vector Embeddings, RAG, Chaining, Structured JSON Output, Function Calling, API calls, Knowledge Graphs, Reinforcement Learning, Fine-tuning, AI Agents. • Why hangout at AI Builders?
Location
We look forward to hang out with you:
207 Old St, London EC1V 9NR, United Kingdom
It's close to the Old Street Tube Station
Thanks to our friends at Beyond
|
AI Builders London Old St. - Mar 5
|
|
AI Builders Lisbon - 4 March
2026-03-04 · 16:00
🎟️ Get Tickets on Lu.ma 🎟️ ☝️This is a donation based meet-up (€5 - €10), Luma ticket is required! Join our technical meetup for the biggest AI nerds in Lisbon! We share the latest AI devtools, APIs, RAG & Agent frameworks, techniques and models. **:: FOR WHO ::** ✅ Anyone actively building with Generative AI ✅ Devs, Product peeps, Data lovers, ML engineers, Founders ⚠️ Basic technical LLM / software engineering knowledge is required!* :: FORMAT ::
**:: AGENDA ::** 17:30 🤝 Walk-in 18:00 🍕 Pizza (be early!) 18:30 🎤 Pioneer Talk 19:00 ---- 💬 BREAK ------ 19:30 💻 Demo 19.45 💻 Demo 20.00 🍻 Drinks 21.00 End ::: FAQ :: • What is AI Builders? A self-organizing nonprofit community of 3000+ AI nerds🤓. and yes.. we're building a democratic AI CEO and run on opencollective.com donations. • Can I demo, give a talk, or just help out? Message Arthur (+31636570260) in case you want to shine✨ on stage and grow your network! • I'm not technical. Can I come? Yes! To fully enjoy the meetup, we recommend chatting with AI to understand these basic LLM concepts: Multimodal, Vector Embeddings, RAG, Chaining, Structured JSON Output, Function Calling, API calls, Knowledge Graphs, Reinforcement Learning, Fine-tuning, AI Agents. • Why hangout at AI Builders?
Why donate? We are a non-profit that relies heavily on donations, pay it forward if you enjoy the events. We are in the process of making our finances totally transparent. If you want to become a recurring supporter for perks, go to : Open Collective. 🎟️ Get Tickets on Lu.ma 🎟️ ☝️This is a donation based meet-up (€5 - €10), Luma ticket is required! |
AI Builders Lisbon - 4 March
|
|
Feb 11 - Visual AI for Video Use Cases
2026-02-11 · 17:00
Join our virtual Meetup to hear talks from experts on cutting-edge topics at the intersection of Visual AI and video use cases. Time and Location Feb 11, 2026 9 - 11 AM Pacific Online. Register for the Zoom! VIDEOP2R: Video Understanding from Perception to Reasoning Reinforcement fine-tuning (RFT), a two-stage framework consisting of supervised fine-tuning (SFT) and reinforcement learning (RL) has shown promising results on improving reasoning ability of large language models (LLMs). Yet extending RFT to large video language models (LVLMs) remains challenging. We propose VideoP2R, a novel process-aware video RFT framework that enhances video reasoning by modeling perception and reasoning as distinct processes. In the SFT stage, we develop a three-step pipeline to generate VideoP2R-CoT-162K, a high-quality, process-aware chain-of-thought (CoT) dataset for perception and reasoning. In the RL stage, we introduce a novel process-aware group relative policy optimization (PA-GRPO) algorithm that supplies separate rewards for perception and reasoning. Extensive experiments show that VideoP2R achieves state-of-the-art (SotA) performance on six out of seven video reasoning and understanding benchmarks. Ablation studies further confirm the effectiveness of our process-aware modeling and PA-GRPO and demonstrate that model's perception output is information-sufficient for downstream reasoning. About the Speaker Yifan Jiang is a third-year Ph.D. student in the Information Science Institute at the University of Southern California (USC-ISI), advised by Dr. Jay Pujara, focusing on natural language processing, commonsense reasoning and multimodality large language models. Layer-Aware Video Composition via Split-then-Merge Split-then-Merge (StM) is a novel generative framework that overcomes data scarcity in video composition by splitting unlabeled videos into separate foreground and background layers for self-supervised learning. By utilizing a transformation-aware training pipeline with multi-layer fusion, the model learns to realistically compose dynamic subjects into diverse scenes without relying on expensive annotated datasets. This presentation will cover the problem of video composition and the details of StM, an approach looking at this problem from a generative AI perspective. We will conclude by demonstrating how StM is working, and outperforming state-of-the-art methods in both quantitative benchmarks and qualitative evaluations. About the Speaker Ozgur Kara is a 4th year Computer Science PhD student at the University of Illinois Urbana-Champaign (UIUC), advised by Founder Professor James M. Rehg. His research builds the next generation of video AI by tackling three core challenges: efficiency, controllability, and safety. Video Reasoning for Worker Safety Ensuring worker safety in industrial environments requires more than object detection or motion tracking; it demands a genuine understanding of human actions, context, and risk. This talk demonstrates how NVIDIA Cosmos Reason, a multimodal video-reasoning model, interprets workplace scenarios with sophisticated temporal and semantic awareness, identifying nuanced safe and unsafe behaviors that conventional vision systems frequently overlook. By integrating Cosmos Reason with FiftyOne, users achieve both automated safety assessments and transparent, interpretable explanations revealing why specific actions are deemed hazardous. Using a curated worker-safety dataset of authentic factory-floor footage, we show how video reasoning enhances audits, training, and compliance workflows while minimizing dependence on extensive labeled datasets. The resulting system demonstrates the potential of explainable multimodal AI to enable safer, more informed decision-making across manufacturing, logistics, construction, healthcare, and other sectors where understanding human behavior is essential. About the Speaker Paula Ramos has a PhD in Computer Vision and Machine Learning, with more than 20 years of experience in the technological field. She has been developing novel integrated engineering technologies, mainly in Computer Vision, robotics, and Machine Learning applied to agriculture, since the early 2000s in Colombia. Video Intelligence Is Going Agentic Video content has become ubiquitous in our digital world, yet the tools for working with video have remained largely unchanged for decades. This talk explores how the convergence of foundation models and agent architectures is fundamentally transforming video interaction and creation. We'll examine how video-native foundation models, multimodal interfaces, and agent transparency are reshaping enterprise media workflows through a deep dive into Jockey, a pioneering video agent system. About the Speaker James Le currently leads the developer experience function at TwelveLabs - a startup building foundation models for video understanding. He previously operated in the MLOps space and ran a blog/podcast on the Data & AI infrastructure ecosystem. |
Feb 11 - Visual AI for Video Use Cases
|
|
Feb 11 - Visual AI for Video Use Cases
2026-02-11 · 17:00
Join our virtual Meetup to hear talks from experts on cutting-edge topics at the intersection of Visual AI and video use cases. Time and Location Feb 11, 2026 9 - 11 AM Pacific Online. Register for the Zoom! VIDEOP2R: Video Understanding from Perception to Reasoning Reinforcement fine-tuning (RFT), a two-stage framework consisting of supervised fine-tuning (SFT) and reinforcement learning (RL) has shown promising results on improving reasoning ability of large language models (LLMs). Yet extending RFT to large video language models (LVLMs) remains challenging. We propose VideoP2R, a novel process-aware video RFT framework that enhances video reasoning by modeling perception and reasoning as distinct processes. In the SFT stage, we develop a three-step pipeline to generate VideoP2R-CoT-162K, a high-quality, process-aware chain-of-thought (CoT) dataset for perception and reasoning. In the RL stage, we introduce a novel process-aware group relative policy optimization (PA-GRPO) algorithm that supplies separate rewards for perception and reasoning. Extensive experiments show that VideoP2R achieves state-of-the-art (SotA) performance on six out of seven video reasoning and understanding benchmarks. Ablation studies further confirm the effectiveness of our process-aware modeling and PA-GRPO and demonstrate that model's perception output is information-sufficient for downstream reasoning. About the Speaker Yifan Jiang is a third-year Ph.D. student in the Information Science Institute at the University of Southern California (USC-ISI), advised by Dr. Jay Pujara, focusing on natural language processing, commonsense reasoning and multimodality large language models. Layer-Aware Video Composition via Split-then-Merge Split-then-Merge (StM) is a novel generative framework that overcomes data scarcity in video composition by splitting unlabeled videos into separate foreground and background layers for self-supervised learning. By utilizing a transformation-aware training pipeline with multi-layer fusion, the model learns to realistically compose dynamic subjects into diverse scenes without relying on expensive annotated datasets. This presentation will cover the problem of video composition and the details of StM, an approach looking at this problem from a generative AI perspective. We will conclude by demonstrating how StM is working, and outperforming state-of-the-art methods in both quantitative benchmarks and qualitative evaluations. About the Speaker Ozgur Kara is a 4th year Computer Science PhD student at the University of Illinois Urbana-Champaign (UIUC), advised by Founder Professor James M. Rehg. His research builds the next generation of video AI by tackling three core challenges: efficiency, controllability, and safety. Video Reasoning for Worker Safety Ensuring worker safety in industrial environments requires more than object detection or motion tracking; it demands a genuine understanding of human actions, context, and risk. This talk demonstrates how NVIDIA Cosmos Reason, a multimodal video-reasoning model, interprets workplace scenarios with sophisticated temporal and semantic awareness, identifying nuanced safe and unsafe behaviors that conventional vision systems frequently overlook. By integrating Cosmos Reason with FiftyOne, users achieve both automated safety assessments and transparent, interpretable explanations revealing why specific actions are deemed hazardous. Using a curated worker-safety dataset of authentic factory-floor footage, we show how video reasoning enhances audits, training, and compliance workflows while minimizing dependence on extensive labeled datasets. The resulting system demonstrates the potential of explainable multimodal AI to enable safer, more informed decision-making across manufacturing, logistics, construction, healthcare, and other sectors where understanding human behavior is essential. About the Speaker Paula Ramos has a PhD in Computer Vision and Machine Learning, with more than 20 years of experience in the technological field. She has been developing novel integrated engineering technologies, mainly in Computer Vision, robotics, and Machine Learning applied to agriculture, since the early 2000s in Colombia. Video Intelligence Is Going Agentic Video content has become ubiquitous in our digital world, yet the tools for working with video have remained largely unchanged for decades. This talk explores how the convergence of foundation models and agent architectures is fundamentally transforming video interaction and creation. We'll examine how video-native foundation models, multimodal interfaces, and agent transparency are reshaping enterprise media workflows through a deep dive into Jockey, a pioneering video agent system. About the Speaker James Le currently leads the developer experience function at TwelveLabs - a startup building foundation models for video understanding. He previously operated in the MLOps space and ran a blog/podcast on the Data & AI infrastructure ecosystem. |
Feb 11 - Visual AI for Video Use Cases
|
|
Feb 11 - Visual AI for Video Use Cases
2026-02-11 · 17:00
Join our virtual Meetup to hear talks from experts on cutting-edge topics at the intersection of Visual AI and video use cases. Time and Location Feb 11, 2026 9 - 11 AM Pacific Online. Register for the Zoom! VIDEOP2R: Video Understanding from Perception to Reasoning Reinforcement fine-tuning (RFT), a two-stage framework consisting of supervised fine-tuning (SFT) and reinforcement learning (RL) has shown promising results on improving reasoning ability of large language models (LLMs). Yet extending RFT to large video language models (LVLMs) remains challenging. We propose VideoP2R, a novel process-aware video RFT framework that enhances video reasoning by modeling perception and reasoning as distinct processes. In the SFT stage, we develop a three-step pipeline to generate VideoP2R-CoT-162K, a high-quality, process-aware chain-of-thought (CoT) dataset for perception and reasoning. In the RL stage, we introduce a novel process-aware group relative policy optimization (PA-GRPO) algorithm that supplies separate rewards for perception and reasoning. Extensive experiments show that VideoP2R achieves state-of-the-art (SotA) performance on six out of seven video reasoning and understanding benchmarks. Ablation studies further confirm the effectiveness of our process-aware modeling and PA-GRPO and demonstrate that model's perception output is information-sufficient for downstream reasoning. About the Speaker Yifan Jiang is a third-year Ph.D. student in the Information Science Institute at the University of Southern California (USC-ISI), advised by Dr. Jay Pujara, focusing on natural language processing, commonsense reasoning and multimodality large language models. Layer-Aware Video Composition via Split-then-Merge Split-then-Merge (StM) is a novel generative framework that overcomes data scarcity in video composition by splitting unlabeled videos into separate foreground and background layers for self-supervised learning. By utilizing a transformation-aware training pipeline with multi-layer fusion, the model learns to realistically compose dynamic subjects into diverse scenes without relying on expensive annotated datasets. This presentation will cover the problem of video composition and the details of StM, an approach looking at this problem from a generative AI perspective. We will conclude by demonstrating how StM is working, and outperforming state-of-the-art methods in both quantitative benchmarks and qualitative evaluations. About the Speaker Ozgur Kara is a 4th year Computer Science PhD student at the University of Illinois Urbana-Champaign (UIUC), advised by Founder Professor James M. Rehg. His research builds the next generation of video AI by tackling three core challenges: efficiency, controllability, and safety. Video Reasoning for Worker Safety Ensuring worker safety in industrial environments requires more than object detection or motion tracking; it demands a genuine understanding of human actions, context, and risk. This talk demonstrates how NVIDIA Cosmos Reason, a multimodal video-reasoning model, interprets workplace scenarios with sophisticated temporal and semantic awareness, identifying nuanced safe and unsafe behaviors that conventional vision systems frequently overlook. By integrating Cosmos Reason with FiftyOne, users achieve both automated safety assessments and transparent, interpretable explanations revealing why specific actions are deemed hazardous. Using a curated worker-safety dataset of authentic factory-floor footage, we show how video reasoning enhances audits, training, and compliance workflows while minimizing dependence on extensive labeled datasets. The resulting system demonstrates the potential of explainable multimodal AI to enable safer, more informed decision-making across manufacturing, logistics, construction, healthcare, and other sectors where understanding human behavior is essential. About the Speaker Paula Ramos has a PhD in Computer Vision and Machine Learning, with more than 20 years of experience in the technological field. She has been developing novel integrated engineering technologies, mainly in Computer Vision, robotics, and Machine Learning applied to agriculture, since the early 2000s in Colombia. Video Intelligence Is Going Agentic Video content has become ubiquitous in our digital world, yet the tools for working with video have remained largely unchanged for decades. This talk explores how the convergence of foundation models and agent architectures is fundamentally transforming video interaction and creation. We'll examine how video-native foundation models, multimodal interfaces, and agent transparency are reshaping enterprise media workflows through a deep dive into Jockey, a pioneering video agent system. About the Speaker James Le currently leads the developer experience function at TwelveLabs - a startup building foundation models for video understanding. He previously operated in the MLOps space and ran a blog/podcast on the Data & AI infrastructure ecosystem. |
Feb 11 - Visual AI for Video Use Cases
|
|
Feb 11 - Visual AI for Video Use Cases
2026-02-11 · 17:00
Join our virtual Meetup to hear talks from experts on cutting-edge topics at the intersection of Visual AI and video use cases. Time and Location Feb 11, 2026 9 - 11 AM Pacific Online. Register for the Zoom! VIDEOP2R: Video Understanding from Perception to Reasoning Reinforcement fine-tuning (RFT), a two-stage framework consisting of supervised fine-tuning (SFT) and reinforcement learning (RL) has shown promising results on improving reasoning ability of large language models (LLMs). Yet extending RFT to large video language models (LVLMs) remains challenging. We propose VideoP2R, a novel process-aware video RFT framework that enhances video reasoning by modeling perception and reasoning as distinct processes. In the SFT stage, we develop a three-step pipeline to generate VideoP2R-CoT-162K, a high-quality, process-aware chain-of-thought (CoT) dataset for perception and reasoning. In the RL stage, we introduce a novel process-aware group relative policy optimization (PA-GRPO) algorithm that supplies separate rewards for perception and reasoning. Extensive experiments show that VideoP2R achieves state-of-the-art (SotA) performance on six out of seven video reasoning and understanding benchmarks. Ablation studies further confirm the effectiveness of our process-aware modeling and PA-GRPO and demonstrate that model's perception output is information-sufficient for downstream reasoning. About the Speaker Yifan Jiang is a third-year Ph.D. student in the Information Science Institute at the University of Southern California (USC-ISI), advised by Dr. Jay Pujara, focusing on natural language processing, commonsense reasoning and multimodality large language models. Layer-Aware Video Composition via Split-then-Merge Split-then-Merge (StM) is a novel generative framework that overcomes data scarcity in video composition by splitting unlabeled videos into separate foreground and background layers for self-supervised learning. By utilizing a transformation-aware training pipeline with multi-layer fusion, the model learns to realistically compose dynamic subjects into diverse scenes without relying on expensive annotated datasets. This presentation will cover the problem of video composition and the details of StM, an approach looking at this problem from a generative AI perspective. We will conclude by demonstrating how StM is working, and outperforming state-of-the-art methods in both quantitative benchmarks and qualitative evaluations. About the Speaker Ozgur Kara is a 4th year Computer Science PhD student at the University of Illinois Urbana-Champaign (UIUC), advised by Founder Professor James M. Rehg. His research builds the next generation of video AI by tackling three core challenges: efficiency, controllability, and safety. Video Reasoning for Worker Safety Ensuring worker safety in industrial environments requires more than object detection or motion tracking; it demands a genuine understanding of human actions, context, and risk. This talk demonstrates how NVIDIA Cosmos Reason, a multimodal video-reasoning model, interprets workplace scenarios with sophisticated temporal and semantic awareness, identifying nuanced safe and unsafe behaviors that conventional vision systems frequently overlook. By integrating Cosmos Reason with FiftyOne, users achieve both automated safety assessments and transparent, interpretable explanations revealing why specific actions are deemed hazardous. Using a curated worker-safety dataset of authentic factory-floor footage, we show how video reasoning enhances audits, training, and compliance workflows while minimizing dependence on extensive labeled datasets. The resulting system demonstrates the potential of explainable multimodal AI to enable safer, more informed decision-making across manufacturing, logistics, construction, healthcare, and other sectors where understanding human behavior is essential. About the Speaker Paula Ramos has a PhD in Computer Vision and Machine Learning, with more than 20 years of experience in the technological field. She has been developing novel integrated engineering technologies, mainly in Computer Vision, robotics, and Machine Learning applied to agriculture, since the early 2000s in Colombia. Video Intelligence Is Going Agentic Video content has become ubiquitous in our digital world, yet the tools for working with video have remained largely unchanged for decades. This talk explores how the convergence of foundation models and agent architectures is fundamentally transforming video interaction and creation. We'll examine how video-native foundation models, multimodal interfaces, and agent transparency are reshaping enterprise media workflows through a deep dive into Jockey, a pioneering video agent system. About the Speaker James Le currently leads the developer experience function at TwelveLabs - a startup building foundation models for video understanding. He previously operated in the MLOps space and ran a blog/podcast on the Data & AI infrastructure ecosystem. |
Feb 11 - Visual AI for Video Use Cases
|
|
Feb 11 - Visual AI for Video Use Cases
2026-02-11 · 17:00
Join our virtual Meetup to hear talks from experts on cutting-edge topics at the intersection of Visual AI and video use cases. Time and Location Feb 11, 2026 9 - 11 AM Pacific Online. Register for the Zoom! VIDEOP2R: Video Understanding from Perception to Reasoning Reinforcement fine-tuning (RFT), a two-stage framework consisting of supervised fine-tuning (SFT) and reinforcement learning (RL) has shown promising results on improving reasoning ability of large language models (LLMs). Yet extending RFT to large video language models (LVLMs) remains challenging. We propose VideoP2R, a novel process-aware video RFT framework that enhances video reasoning by modeling perception and reasoning as distinct processes. In the SFT stage, we develop a three-step pipeline to generate VideoP2R-CoT-162K, a high-quality, process-aware chain-of-thought (CoT) dataset for perception and reasoning. In the RL stage, we introduce a novel process-aware group relative policy optimization (PA-GRPO) algorithm that supplies separate rewards for perception and reasoning. Extensive experiments show that VideoP2R achieves state-of-the-art (SotA) performance on six out of seven video reasoning and understanding benchmarks. Ablation studies further confirm the effectiveness of our process-aware modeling and PA-GRPO and demonstrate that model's perception output is information-sufficient for downstream reasoning. About the Speaker Yifan Jiang is a third-year Ph.D. student in the Information Science Institute at the University of Southern California (USC-ISI), advised by Dr. Jay Pujara, focusing on natural language processing, commonsense reasoning and multimodality large language models. Layer-Aware Video Composition via Split-then-Merge Split-then-Merge (StM) is a novel generative framework that overcomes data scarcity in video composition by splitting unlabeled videos into separate foreground and background layers for self-supervised learning. By utilizing a transformation-aware training pipeline with multi-layer fusion, the model learns to realistically compose dynamic subjects into diverse scenes without relying on expensive annotated datasets. This presentation will cover the problem of video composition and the details of StM, an approach looking at this problem from a generative AI perspective. We will conclude by demonstrating how StM is working, and outperforming state-of-the-art methods in both quantitative benchmarks and qualitative evaluations. About the Speaker Ozgur Kara is a 4th year Computer Science PhD student at the University of Illinois Urbana-Champaign (UIUC), advised by Founder Professor James M. Rehg. His research builds the next generation of video AI by tackling three core challenges: efficiency, controllability, and safety. Video Reasoning for Worker Safety Ensuring worker safety in industrial environments requires more than object detection or motion tracking; it demands a genuine understanding of human actions, context, and risk. This talk demonstrates how NVIDIA Cosmos Reason, a multimodal video-reasoning model, interprets workplace scenarios with sophisticated temporal and semantic awareness, identifying nuanced safe and unsafe behaviors that conventional vision systems frequently overlook. By integrating Cosmos Reason with FiftyOne, users achieve both automated safety assessments and transparent, interpretable explanations revealing why specific actions are deemed hazardous. Using a curated worker-safety dataset of authentic factory-floor footage, we show how video reasoning enhances audits, training, and compliance workflows while minimizing dependence on extensive labeled datasets. The resulting system demonstrates the potential of explainable multimodal AI to enable safer, more informed decision-making across manufacturing, logistics, construction, healthcare, and other sectors where understanding human behavior is essential. About the Speaker Paula Ramos has a PhD in Computer Vision and Machine Learning, with more than 20 years of experience in the technological field. She has been developing novel integrated engineering technologies, mainly in Computer Vision, robotics, and Machine Learning applied to agriculture, since the early 2000s in Colombia. Video Intelligence Is Going Agentic Video content has become ubiquitous in our digital world, yet the tools for working with video have remained largely unchanged for decades. This talk explores how the convergence of foundation models and agent architectures is fundamentally transforming video interaction and creation. We'll examine how video-native foundation models, multimodal interfaces, and agent transparency are reshaping enterprise media workflows through a deep dive into Jockey, a pioneering video agent system. About the Speaker James Le currently leads the developer experience function at TwelveLabs - a startup building foundation models for video understanding. He previously operated in the MLOps space and ran a blog/podcast on the Data & AI infrastructure ecosystem. |
Feb 11 - Visual AI for Video Use Cases
|
|
Feb 11 - Visual AI for Video Use Cases
2026-02-11 · 17:00
Join our virtual Meetup to hear talks from experts on cutting-edge topics at the intersection of Visual AI and video use cases. Time and Location Feb 11, 2026 9 - 11 AM Pacific Online. Register for the Zoom! VIDEOP2R: Video Understanding from Perception to Reasoning Reinforcement fine-tuning (RFT), a two-stage framework consisting of supervised fine-tuning (SFT) and reinforcement learning (RL) has shown promising results on improving reasoning ability of large language models (LLMs). Yet extending RFT to large video language models (LVLMs) remains challenging. We propose VideoP2R, a novel process-aware video RFT framework that enhances video reasoning by modeling perception and reasoning as distinct processes. In the SFT stage, we develop a three-step pipeline to generate VideoP2R-CoT-162K, a high-quality, process-aware chain-of-thought (CoT) dataset for perception and reasoning. In the RL stage, we introduce a novel process-aware group relative policy optimization (PA-GRPO) algorithm that supplies separate rewards for perception and reasoning. Extensive experiments show that VideoP2R achieves state-of-the-art (SotA) performance on six out of seven video reasoning and understanding benchmarks. Ablation studies further confirm the effectiveness of our process-aware modeling and PA-GRPO and demonstrate that model's perception output is information-sufficient for downstream reasoning. About the Speaker Yifan Jiang is a third-year Ph.D. student in the Information Science Institute at the University of Southern California (USC-ISI), advised by Dr. Jay Pujara, focusing on natural language processing, commonsense reasoning and multimodality large language models. Layer-Aware Video Composition via Split-then-Merge Split-then-Merge (StM) is a novel generative framework that overcomes data scarcity in video composition by splitting unlabeled videos into separate foreground and background layers for self-supervised learning. By utilizing a transformation-aware training pipeline with multi-layer fusion, the model learns to realistically compose dynamic subjects into diverse scenes without relying on expensive annotated datasets. This presentation will cover the problem of video composition and the details of StM, an approach looking at this problem from a generative AI perspective. We will conclude by demonstrating how StM is working, and outperforming state-of-the-art methods in both quantitative benchmarks and qualitative evaluations. About the Speaker Ozgur Kara is a 4th year Computer Science PhD student at the University of Illinois Urbana-Champaign (UIUC), advised by Founder Professor James M. Rehg. His research builds the next generation of video AI by tackling three core challenges: efficiency, controllability, and safety. Video Reasoning for Worker Safety Ensuring worker safety in industrial environments requires more than object detection or motion tracking; it demands a genuine understanding of human actions, context, and risk. This talk demonstrates how NVIDIA Cosmos Reason, a multimodal video-reasoning model, interprets workplace scenarios with sophisticated temporal and semantic awareness, identifying nuanced safe and unsafe behaviors that conventional vision systems frequently overlook. By integrating Cosmos Reason with FiftyOne, users achieve both automated safety assessments and transparent, interpretable explanations revealing why specific actions are deemed hazardous. Using a curated worker-safety dataset of authentic factory-floor footage, we show how video reasoning enhances audits, training, and compliance workflows while minimizing dependence on extensive labeled datasets. The resulting system demonstrates the potential of explainable multimodal AI to enable safer, more informed decision-making across manufacturing, logistics, construction, healthcare, and other sectors where understanding human behavior is essential. About the Speaker Paula Ramos has a PhD in Computer Vision and Machine Learning, with more than 20 years of experience in the technological field. She has been developing novel integrated engineering technologies, mainly in Computer Vision, robotics, and Machine Learning applied to agriculture, since the early 2000s in Colombia. Video Intelligence Is Going Agentic Video content has become ubiquitous in our digital world, yet the tools for working with video have remained largely unchanged for decades. This talk explores how the convergence of foundation models and agent architectures is fundamentally transforming video interaction and creation. We'll examine how video-native foundation models, multimodal interfaces, and agent transparency are reshaping enterprise media workflows through a deep dive into Jockey, a pioneering video agent system. About the Speaker James Le currently leads the developer experience function at TwelveLabs - a startup building foundation models for video understanding. He previously operated in the MLOps space and ran a blog/podcast on the Data & AI infrastructure ecosystem. |
Feb 11 - Visual AI for Video Use Cases
|
|
Feb 11 - Visual AI for Video Use Cases
2026-02-11 · 17:00
Join our virtual Meetup to hear talks from experts on cutting-edge topics at the intersection of Visual AI and video use cases. Time and Location Feb 11, 2026 9 - 11 AM Pacific Online. Register for the Zoom! VIDEOP2R: Video Understanding from Perception to Reasoning Reinforcement fine-tuning (RFT), a two-stage framework consisting of supervised fine-tuning (SFT) and reinforcement learning (RL) has shown promising results on improving reasoning ability of large language models (LLMs). Yet extending RFT to large video language models (LVLMs) remains challenging. We propose VideoP2R, a novel process-aware video RFT framework that enhances video reasoning by modeling perception and reasoning as distinct processes. In the SFT stage, we develop a three-step pipeline to generate VideoP2R-CoT-162K, a high-quality, process-aware chain-of-thought (CoT) dataset for perception and reasoning. In the RL stage, we introduce a novel process-aware group relative policy optimization (PA-GRPO) algorithm that supplies separate rewards for perception and reasoning. Extensive experiments show that VideoP2R achieves state-of-the-art (SotA) performance on six out of seven video reasoning and understanding benchmarks. Ablation studies further confirm the effectiveness of our process-aware modeling and PA-GRPO and demonstrate that model's perception output is information-sufficient for downstream reasoning. About the Speaker Yifan Jiang is a third-year Ph.D. student in the Information Science Institute at the University of Southern California (USC-ISI), advised by Dr. Jay Pujara, focusing on natural language processing, commonsense reasoning and multimodality large language models. Layer-Aware Video Composition via Split-then-Merge Split-then-Merge (StM) is a novel generative framework that overcomes data scarcity in video composition by splitting unlabeled videos into separate foreground and background layers for self-supervised learning. By utilizing a transformation-aware training pipeline with multi-layer fusion, the model learns to realistically compose dynamic subjects into diverse scenes without relying on expensive annotated datasets. This presentation will cover the problem of video composition and the details of StM, an approach looking at this problem from a generative AI perspective. We will conclude by demonstrating how StM is working, and outperforming state-of-the-art methods in both quantitative benchmarks and qualitative evaluations. About the Speaker Ozgur Kara is a 4th year Computer Science PhD student at the University of Illinois Urbana-Champaign (UIUC), advised by Founder Professor James M. Rehg. His research builds the next generation of video AI by tackling three core challenges: efficiency, controllability, and safety. Video Reasoning for Worker Safety Ensuring worker safety in industrial environments requires more than object detection or motion tracking; it demands a genuine understanding of human actions, context, and risk. This talk demonstrates how NVIDIA Cosmos Reason, a multimodal video-reasoning model, interprets workplace scenarios with sophisticated temporal and semantic awareness, identifying nuanced safe and unsafe behaviors that conventional vision systems frequently overlook. By integrating Cosmos Reason with FiftyOne, users achieve both automated safety assessments and transparent, interpretable explanations revealing why specific actions are deemed hazardous. Using a curated worker-safety dataset of authentic factory-floor footage, we show how video reasoning enhances audits, training, and compliance workflows while minimizing dependence on extensive labeled datasets. The resulting system demonstrates the potential of explainable multimodal AI to enable safer, more informed decision-making across manufacturing, logistics, construction, healthcare, and other sectors where understanding human behavior is essential. About the Speaker Paula Ramos has a PhD in Computer Vision and Machine Learning, with more than 20 years of experience in the technological field. She has been developing novel integrated engineering technologies, mainly in Computer Vision, robotics, and Machine Learning applied to agriculture, since the early 2000s in Colombia. Video Intelligence Is Going Agentic Video content has become ubiquitous in our digital world, yet the tools for working with video have remained largely unchanged for decades. This talk explores how the convergence of foundation models and agent architectures is fundamentally transforming video interaction and creation. We'll examine how video-native foundation models, multimodal interfaces, and agent transparency are reshaping enterprise media workflows through a deep dive into Jockey, a pioneering video agent system. About the Speaker James Le currently leads the developer experience function at TwelveLabs - a startup building foundation models for video understanding. He previously operated in the MLOps space and ran a blog/podcast on the Data & AI infrastructure ecosystem. |
Feb 11 - Visual AI for Video Use Cases
|
|
Feb 11 - Visual AI for Video Use Cases
2026-02-11 · 17:00
Join our virtual Meetup to hear talks from experts on cutting-edge topics at the intersection of Visual AI and video use cases. Time and Location Feb 11, 2026 9 - 11 AM Pacific Online. Register for the Zoom! VIDEOP2R: Video Understanding from Perception to Reasoning Reinforcement fine-tuning (RFT), a two-stage framework consisting of supervised fine-tuning (SFT) and reinforcement learning (RL) has shown promising results on improving reasoning ability of large language models (LLMs). Yet extending RFT to large video language models (LVLMs) remains challenging. We propose VideoP2R, a novel process-aware video RFT framework that enhances video reasoning by modeling perception and reasoning as distinct processes. In the SFT stage, we develop a three-step pipeline to generate VideoP2R-CoT-162K, a high-quality, process-aware chain-of-thought (CoT) dataset for perception and reasoning. In the RL stage, we introduce a novel process-aware group relative policy optimization (PA-GRPO) algorithm that supplies separate rewards for perception and reasoning. Extensive experiments show that VideoP2R achieves state-of-the-art (SotA) performance on six out of seven video reasoning and understanding benchmarks. Ablation studies further confirm the effectiveness of our process-aware modeling and PA-GRPO and demonstrate that model's perception output is information-sufficient for downstream reasoning. About the Speaker Yifan Jiang is a third-year Ph.D. student in the Information Science Institute at the University of Southern California (USC-ISI), advised by Dr. Jay Pujara, focusing on natural language processing, commonsense reasoning and multimodality large language models. Layer-Aware Video Composition via Split-then-Merge Split-then-Merge (StM) is a novel generative framework that overcomes data scarcity in video composition by splitting unlabeled videos into separate foreground and background layers for self-supervised learning. By utilizing a transformation-aware training pipeline with multi-layer fusion, the model learns to realistically compose dynamic subjects into diverse scenes without relying on expensive annotated datasets. This presentation will cover the problem of video composition and the details of StM, an approach looking at this problem from a generative AI perspective. We will conclude by demonstrating how StM is working, and outperforming state-of-the-art methods in both quantitative benchmarks and qualitative evaluations. About the Speaker Ozgur Kara is a 4th year Computer Science PhD student at the University of Illinois Urbana-Champaign (UIUC), advised by Founder Professor James M. Rehg. His research builds the next generation of video AI by tackling three core challenges: efficiency, controllability, and safety. Video Reasoning for Worker Safety Ensuring worker safety in industrial environments requires more than object detection or motion tracking; it demands a genuine understanding of human actions, context, and risk. This talk demonstrates how NVIDIA Cosmos Reason, a multimodal video-reasoning model, interprets workplace scenarios with sophisticated temporal and semantic awareness, identifying nuanced safe and unsafe behaviors that conventional vision systems frequently overlook. By integrating Cosmos Reason with FiftyOne, users achieve both automated safety assessments and transparent, interpretable explanations revealing why specific actions are deemed hazardous. Using a curated worker-safety dataset of authentic factory-floor footage, we show how video reasoning enhances audits, training, and compliance workflows while minimizing dependence on extensive labeled datasets. The resulting system demonstrates the potential of explainable multimodal AI to enable safer, more informed decision-making across manufacturing, logistics, construction, healthcare, and other sectors where understanding human behavior is essential. About the Speaker Paula Ramos has a PhD in Computer Vision and Machine Learning, with more than 20 years of experience in the technological field. She has been developing novel integrated engineering technologies, mainly in Computer Vision, robotics, and Machine Learning applied to agriculture, since the early 2000s in Colombia. Video Intelligence Is Going Agentic Video content has become ubiquitous in our digital world, yet the tools for working with video have remained largely unchanged for decades. This talk explores how the convergence of foundation models and agent architectures is fundamentally transforming video interaction and creation. We'll examine how video-native foundation models, multimodal interfaces, and agent transparency are reshaping enterprise media workflows through a deep dive into Jockey, a pioneering video agent system. About the Speaker James Le currently leads the developer experience function at TwelveLabs - a startup building foundation models for video understanding. He previously operated in the MLOps space and ran a blog/podcast on the Data & AI infrastructure ecosystem. |
Feb 11 - Visual AI for Video Use Cases
|
|
AI Builders London Old St. - Feb 5
2026-02-05 · 17:30
🎟️ Get tickets: https://lu.ma/ai-builders 🎟️ ☝️This is a free meetup however a Luma ticket is required! Join our meetup for AI engineers & founders! We share the latest insights about: AI dev tools, Agent frameworks, RAG pipelines, automation hacks, Cursor/Claude code hacks, and new gen-AI models. **:: FOR WHO ::** ✅ Anyone actively building with Generative AI ✅ Devs, Product peeps, Data lovers, ML engineers, Founders ⚠️Technical LLM knowledge required!* :: FORMAT ::
**:: AGENDA ::** 17:30 🤝 Walk-in 18:00 🍕 Pizza (be early!) 18:30 🎤 Pioneer Speaker 19:00 ---- 💬 BREAK TIME ----- 19:30 💻 Demo (TBA) 19.45 💻 Demo (TBA) 20.00 🍻 Drinks 21.00 End :: FAQ :: • What is AI Builders? A self-organizing nonprofit community of 3000+ AI nerds🤓. and yes.. we're building a democratic AI CEO and run on opencollective.com donations. • Can I demo, give a talk, or just help out? Message Arthur (+31636570260) in case you want to shine✨ on stage and grow your network! • *I'm not technical. Can I come? Yes! To fully enjoy the meetup, we recommend chatting with AI to understand these basic LLM concepts: Multimodal, Vector Embeddings, RAG, Chaining, Structured JSON Output, Function Calling, API calls, Knowledge Graphs, Reinforcement Learning, Fine-tuning, AI Agents. • Why hangout at AI Builders?
Location
We look forward to hang out with you:
207 Old St, London EC1V 9NR, United Kingdom
It's close to the Old Street Tube Station
Thanks to our friends at Beyond
|
AI Builders London Old St. - Feb 5
|
|
AI Builders Lisbon - 4 February
2026-02-04 · 16:00
🎟️ Get Tickets on Lu.ma 🎟️ ☝️This is a donation based meet-up (€5 - €10), Luma ticket is required! Join our technical meetup for the biggest AI nerds in Lisbon! We share the latest AI devtools, APIs, RAG & Agent frameworks, techniques and models. **:: FOR WHO ::** ✅ Anyone actively building with Generative AI ✅ Devs, Product peeps, Data lovers, ML engineers, Founders ⚠️ Basic technical LLM / software engineering knowledge is required!* :: FORMAT ::
**:: AGENDA ::** 17:30 🤝 Walk-in 18:00 🍕 Pizza (be early!) 18:30 🎤 Pioneer Talk 19:00 ---- 💬 BREAK ------ 19:30 💻 Demo 19.45 💻 Demo 20.00 🍻 Drinks 21.00 End ::: FAQ :: • What is AI Builders? A self-organizing nonprofit community of 3000+ AI nerds🤓. and yes.. we're building a democratic AI CEO and run on opencollective.com donations. • Can I demo, give a talk, or just help out? Message Arthur (+31636570260) in case you want to shine✨ on stage and grow your network! • I'm not technical. Can I come? Yes! To fully enjoy the meetup, we recommend chatting with AI to understand these basic LLM concepts: Multimodal, Vector Embeddings, RAG, Chaining, Structured JSON Output, Function Calling, API calls, Knowledge Graphs, Reinforcement Learning, Fine-tuning, AI Agents. • Why hangout at AI Builders?
Why donate? We are a non-profit that relies heavily on donations, pay it forward if you enjoy the events. We are in the process of making our finances totally transparent. If you want to become a recurring supporter for perks, go to : Open Collective. 🎟️ Get Tickets on Lu.ma 🎟️ ☝️This is a donation based meet-up (€5 - €10), Luma ticket is required! |
AI Builders Lisbon - 4 February
|
|
AWS x Datadog re:Invent Recap
2026-01-29 · 16:45
Welcome to the first Datadog User Group Meetup of the year! We’re kicking things off with an AWS re:Invent Recap on Thursday January 29th, bringing the community together to share insights, highlights and practical takeaways from the conference. This year’s agenda also puts a strong spotlight on AI and how it is rapidly reshaping DevOps, operations and on-call workflows, setting the stage for an exciting year ahead. That's also why we've invited Prosus to the stage, leading the way in the AI field, focusing heavily on integrating artificial intelligence into its global e-commerce and lifestyle brands, as well as investing in the broader AI ecosystem. Looking forward to seeing you there and starting the year strong as a local Datadog & AWS community. Agenda 17:30 Walk-in, food and drinks (vegetarian options available) 18:00 Opening by Datadog User Group Organizers --- 18.10 Talk 1: AWS DevOps Agent, drive operational excellence with a frontier agent that resolves and proactively prevents incidents by Ioannis Moustakis (Solutions Architect Lead, AWS) 18.50 Talk 2: TBD by NAME (TITLE, Prosus) 19.30 Quick Break 19.40 Talk 3: Introducing Bits AI SRE, your AI on-call teammate by Ryan Earley (Sr. Enterprise Sales Engineer, Datadog) --- 20:20 Drinks and socializing 21:00 - 21:15 End Speaker Bio's: 1. Ioannis Moustakis\, Solutions Architecture Lead (AWS) Leading a team of Solutions Architects across 11 countries in Northern Europe, supporting mid-market and scale-ups to accelerate cloud and AI adoption, modernization, and measurable business outcomes on AWS. 2. NAME\, Title (Prosus) Speaker Bio.. 3. Ryan Earley\, Sr. Enterprise Sales Engineer (Datadog) Ryan has been with Datadog for over 4,5 years, supporting Enteprrise organizations in EMEA with the evaluation and adoption of AI Observability and Security practices. Privacy notice(s):
Other important notes:
|
AWS x Datadog re:Invent Recap
|
|
AI Webinar Series (Virtual) - Evaluating AI Agent Reliability
2026-01-21 · 18:00
Important: Register on the event website to receive joining link. (rsvp on meetup will NOT receive joining link). This is virtual event for our AI global community, please double-check your local time. Can't make it live? Register anyway to receive the webinar recording. Description: Welcome to the weekly AI Deep Dive Webinar Series. Join us for deep dive tech talks on AI, hands-on experiences on code labs, workshops, and networking with speakers & fellow developers from all over the world. Tech Talk: Evaluating AI Agent Reliability Speaker: Anupam Datta (Snowflake) \| Josh Reini (Snowflake) Abstract: Agents often fail in ways you can’t see. They could return a final answer while taking a broken path: drifting from the goal, making irrational plan jumps, or misusing tools. Was the goal achieved efficiently? Did the plan make sense? Were the right tools used? Did the agent follow through? These hidden mistakes silently rack up compute costs, spike latency, and cause brittle behavior that collapses in production. Traditional evals won’t flag any of it because they only check the output, not the decisions that produced it. This session introduces the Agent GPA (Goal-Plan-Action) framework, available in the open-source TruLens library. Benchmark tests show the Agent GPA framework consistently outperformed standard LLM evaluators, giving teams scalable and trustworthy insight into agent behavior
You’ll learn how to inspect an agent’s reasoning steps, detect issues like hallucinations, bad tool calls, and missed actions, and leave knowing how to make your agent truly production-ready. Speakers/Topics: Stay tuned as we are updating speakers and schedules. If you have a keen interest in speaking to our community, we invite you to submit topics for consideration: Submit Topics More upcoming sessions:
Local and Global AI Community on Discord Join us on discord for local and global AI tech community:
|
AI Webinar Series (Virtual) - Evaluating AI Agent Reliability
|
|
AI Webinar Series (Virtual) - Evaluating AI Agent Reliability
2026-01-21 · 18:00
Important: Register on the event website to receive joining link. (rsvp on meetup will NOT receive joining link). This is virtual event for our AI global community, please double-check your local time. Can't make it live? Register anyway to receive the webinar recording. Description: Welcome to the weekly AI Deep Dive Webinar Series. Join us for deep dive tech talks on AI, hands-on experiences on code labs, workshops, and networking with speakers & fellow developers from all over the world. Tech Talk: Evaluating AI Agent Reliability Speaker: Anupam Datta (Snowflake) \| Josh Reini (Snowflake) Abstract: Agents often fail in ways you can’t see. They could return a final answer while taking a broken path: drifting from the goal, making irrational plan jumps, or misusing tools. Was the goal achieved efficiently? Did the plan make sense? Were the right tools used? Did the agent follow through? These hidden mistakes silently rack up compute costs, spike latency, and cause brittle behavior that collapses in production. Traditional evals won’t flag any of it because they only check the output, not the decisions that produced it. This session introduces the Agent GPA (Goal-Plan-Action) framework, available in the open-source TruLens library. Benchmark tests show the Agent GPA framework consistently outperformed standard LLM evaluators, giving teams scalable and trustworthy insight into agent behavior
You’ll learn how to inspect an agent’s reasoning steps, detect issues like hallucinations, bad tool calls, and missed actions, and leave knowing how to make your agent truly production-ready. Speakers/Topics: Stay tuned as we are updating speakers and schedules. If you have a keen interest in speaking to our community, we invite you to submit topics for consideration: Submit Topics More upcoming sessions:
Local and Global AI Community on Discord Join us on discord for local and global AI tech community:
|
AI Webinar Series (Virtual) - Evaluating AI Agent Reliability
|
|
AI Webinar Series (Virtual) - Evaluating AI Agent Reliability
2026-01-21 · 18:00
Important: Register on the event website to receive joining link. (rsvp on meetup will NOT receive joining link). This is virtual event for our AI global community, please double-check your local time. Can't make it live? Register anyway to receive the webinar recording. Description: Welcome to the weekly AI Deep Dive Webinar Series. Join us for deep dive tech talks on AI, hands-on experiences on code labs, workshops, and networking with speakers & fellow developers from all over the world. Tech Talk: Evaluating AI Agent Reliability Speaker: Anupam Datta (Snowflake) \| Josh Reini (Snowflake) Abstract: Agents often fail in ways you can’t see. They could return a final answer while taking a broken path: drifting from the goal, making irrational plan jumps, or misusing tools. Was the goal achieved efficiently? Did the plan make sense? Were the right tools used? Did the agent follow through? These hidden mistakes silently rack up compute costs, spike latency, and cause brittle behavior that collapses in production. Traditional evals won’t flag any of it because they only check the output, not the decisions that produced it. This session introduces the Agent GPA (Goal-Plan-Action) framework, available in the open-source TruLens library. Benchmark tests show the Agent GPA framework consistently outperformed standard LLM evaluators, giving teams scalable and trustworthy insight into agent behavior
You’ll learn how to inspect an agent’s reasoning steps, detect issues like hallucinations, bad tool calls, and missed actions, and leave knowing how to make your agent truly production-ready. Speakers/Topics: Stay tuned as we are updating speakers and schedules. If you have a keen interest in speaking to our community, we invite you to submit topics for consideration: Submit Topics More upcoming sessions:
Local and Global AI Community on Discord Join us on discord for local and global AI tech community:
|
AI Webinar Series (Virtual) - Evaluating AI Agent Reliability
|
|
AI Webinar Series (Virtual) - Evaluating AI Agent Reliability
2026-01-21 · 18:00
Important: Register on the event website to receive joining link. (rsvp on meetup will NOT receive joining link). This is virtual event for our AI global community, please double-check your local time. Can't make it live? Register anyway to receive the webinar recording. Description: Welcome to the weekly AI Deep Dive Webinar Series. Join us for deep dive tech talks on AI, hands-on experiences on code labs, workshops, and networking with speakers & fellow developers from all over the world. Tech Talk: Evaluating AI Agent Reliability Speaker: Anupam Datta (Snowflake) \| Josh Reini (Snowflake) Abstract: Agents often fail in ways you can’t see. They could return a final answer while taking a broken path: drifting from the goal, making irrational plan jumps, or misusing tools. Was the goal achieved efficiently? Did the plan make sense? Were the right tools used? Did the agent follow through? These hidden mistakes silently rack up compute costs, spike latency, and cause brittle behavior that collapses in production. Traditional evals won’t flag any of it because they only check the output, not the decisions that produced it. This session introduces the Agent GPA (Goal-Plan-Action) framework, available in the open-source TruLens library. Benchmark tests show the Agent GPA framework consistently outperformed standard LLM evaluators, giving teams scalable and trustworthy insight into agent behavior
You’ll learn how to inspect an agent’s reasoning steps, detect issues like hallucinations, bad tool calls, and missed actions, and leave knowing how to make your agent truly production-ready. Speakers/Topics: Stay tuned as we are updating speakers and schedules. If you have a keen interest in speaking to our community, we invite you to submit topics for consideration: Submit Topics More upcoming sessions:
Local and Global AI Community on Discord Join us on discord for local and global AI tech community:
|
AI Webinar Series (Virtual) - Evaluating AI Agent Reliability
|
|
AI Webinar Series - Evaluating AI Agent Reliability
2026-01-21 · 18:00
Important: Register on the event website to receive joining link. (rsvp on meetup will NOT receive joining link). This is virtual event for our AI global community, please double-check your local time. Can't make it live? Register anyway to receive the webinar recording. Description: Welcome to the weekly AI Deep Dive Webinar Series. Join us for deep dive tech talks on AI, hands-on experiences on code labs, workshops, and networking with speakers & fellow developers from all over the world. Tech Talk: Evaluating AI Agent Reliability Speaker: Anupam Datta (Snowflake) \| Josh Reini (Snowflake) Abstract: Agents often fail in ways you can’t see. They could return a final answer while taking a broken path: drifting from the goal, making irrational plan jumps, or misusing tools. Was the goal achieved efficiently? Did the plan make sense? Were the right tools used? Did the agent follow through? These hidden mistakes silently rack up compute costs, spike latency, and cause brittle behavior that collapses in production. Traditional evals won’t flag any of it because they only check the output, not the decisions that produced it. This session introduces the Agent GPA (Goal-Plan-Action) framework, available in the open-source TruLens library. Benchmark tests show the Agent GPA framework consistently outperformed standard LLM evaluators, giving teams scalable and trustworthy insight into agent behavior
You’ll learn how to inspect an agent’s reasoning steps, detect issues like hallucinations, bad tool calls, and missed actions, and leave knowing how to make your agent truly production-ready. Speakers/Topics: Stay tuned as we are updating speakers and schedules. If you have a keen interest in speaking to our community, we invite you to submit topics for consideration: Submit Topics More upcoming sessions:
Local and Global AI Community on Discord Join us on discord for local and global AI tech community:
|
AI Webinar Series - Evaluating AI Agent Reliability
|
|
AI Webinar Series - Evaluating AI Agent Reliability
2026-01-21 · 18:00
Important: Register on the event website to receive joining link. (rsvp on meetup will NOT receive joining link). This is virtual event for our AI global community, please double-check your local time. Can't make it live? Register anyway to receive the webinar recording. Description: Welcome to the weekly AI Deep Dive Webinar Series. Join us for deep dive tech talks on AI, hands-on experiences on code labs, workshops, and networking with speakers & fellow developers from all over the world. Tech Talk: Evaluating AI Agent Reliability Speaker: Anupam Datta (Snowflake) \| Josh Reini (Snowflake) Abstract: Agents often fail in ways you can’t see. They could return a final answer while taking a broken path: drifting from the goal, making irrational plan jumps, or misusing tools. Was the goal achieved efficiently? Did the plan make sense? Were the right tools used? Did the agent follow through? These hidden mistakes silently rack up compute costs, spike latency, and cause brittle behavior that collapses in production. Traditional evals won’t flag any of it because they only check the output, not the decisions that produced it. This session introduces the Agent GPA (Goal-Plan-Action) framework, available in the open-source TruLens library. Benchmark tests show the Agent GPA framework consistently outperformed standard LLM evaluators, giving teams scalable and trustworthy insight into agent behavior
You’ll learn how to inspect an agent’s reasoning steps, detect issues like hallucinations, bad tool calls, and missed actions, and leave knowing how to make your agent truly production-ready. Speakers/Topics: Stay tuned as we are updating speakers and schedules. If you have a keen interest in speaking to our community, we invite you to submit topics for consideration: Submit Topics More upcoming sessions:
Local and Global AI Community on Discord Join us on discord for local and global AI tech community:
|
AI Webinar Series - Evaluating AI Agent Reliability
|
|
AI Webinar Series (Virtual) - Evaluating AI Agent Reliability
2026-01-21 · 18:00
Important: Register on the event website to receive joining link. (rsvp on meetup will NOT receive joining link). This is virtual event for our AI global community, please double-check your local time. Can't make it live? Register anyway to receive the webinar recording. Description: Welcome to the weekly AI Deep Dive Webinar Series. Join us for deep dive tech talks on AI, hands-on experiences on code labs, workshops, and networking with speakers & fellow developers from all over the world. Tech Talk: Evaluating AI Agent Reliability Speaker: Anupam Datta (Snowflake) \| Josh Reini (Snowflake) Abstract: Agents often fail in ways you can’t see. They could return a final answer while taking a broken path: drifting from the goal, making irrational plan jumps, or misusing tools. Was the goal achieved efficiently? Did the plan make sense? Were the right tools used? Did the agent follow through? These hidden mistakes silently rack up compute costs, spike latency, and cause brittle behavior that collapses in production. Traditional evals won’t flag any of it because they only check the output, not the decisions that produced it. This session introduces the Agent GPA (Goal-Plan-Action) framework, available in the open-source TruLens library. Benchmark tests show the Agent GPA framework consistently outperformed standard LLM evaluators, giving teams scalable and trustworthy insight into agent behavior
You’ll learn how to inspect an agent’s reasoning steps, detect issues like hallucinations, bad tool calls, and missed actions, and leave knowing how to make your agent truly production-ready. Speakers/Topics: Stay tuned as we are updating speakers and schedules. If you have a keen interest in speaking to our community, we invite you to submit topics for consideration: Submit Topics More upcoming sessions:
Local and Global AI Community on Discord Join us on discord for local and global AI tech community:
|
AI Webinar Series (Virtual) - Evaluating AI Agent Reliability
|