talk-data.com

We're excited to invite you to join us on 18 September to for our latest London event: THE AI SHIFT: Transforming the way we Buy, Run and Fund Cloud.

We all know AI tools and agents are changing the cloud game and in this event we'll explore how we can shape the future with Ai, rather than being shaped by it.

For those of you who were at our last London event, 'Fun with FinOps' on 12 June, we're using the same formula of mixing fun, sponsor-hosted games with practitioner talks, all on the theme of how AI is changing how we work with cloud.

Please note: RSVP on Meetup is not valid for entry t this event.

To attend “THE AI SHIFT – Transforming How We Build, Run & Fund Cloud” on 18 September, you must register through our official portal: Register here to secure your place

This helps us personalise the experience, assign teams for the interactive games, and make sure you're fully in the loop. Spots are limited — don’t miss out!

We’ll email you full event details once your're registered! Mike Bradbury Group organizer

We're excited to invite you to join us on 18 September to for our latest London event: THE AI SHIFT: Transforming the way we Buy, Run and Fund Cloud.

We all know AI tools and agents are changing the cloud game and in this event we'll explore how we can shape the future with Ai, rather than being shaped by it.

For those of you who were at our last London event, 'Fun with FinOps' on 12 June, we're using the same formula of mixing fun, sponsor-hosted games with practitioner talks, all on the theme of how AI is changing how we work with cloud.

Please note: RSVP on Meetup is not valid for entry t this event.

To attend “THE AI SHIFT – Transforming How We Build, Run & Fund Cloud” on 18 September, you must register through our official portal: Register here to secure your place

This helps us personalise the experience, assign teams for the interactive games, and make sure you're fully in the loop. Spots are limited — don’t miss out!

We’ll email you full event details once your're registered! Mike Bradbury Group organizer

We're excited to invite you to join us on 18 September to for our latest London event: THE AI SHIFT: Transforming the way we Buy, Run and Fund Cloud.

We all know AI tools and agents are changing the cloud game and in this event we'll explore how we can shape the future with Ai, rather than being shaped by it.

For those of you who were at our last London event, 'Fun with FinOps' on 12 June, we're using the same formula of mixing fun, sponsor-hosted games with practitioner talks, all on the theme of how AI is changing how we work with cloud.

Please note: RSVP on Meetup is not valid for entry t this event.

To attend “THE AI SHIFT – Transforming How We Build, Run & Fund Cloud” on 18 September, you must register through our official portal: Register here to secure your place

This helps us personalise the experience, assign teams for the interactive games, and make sure you're fully in the loop. Spots are limited — don’t miss out!

We’ll email you full event details once your're registered! Mike Bradbury Group organizer

Venue: Carnival House, 100 Harbour Parade, Southampton, SO15 1ST 📢 Want to speak 📢: submit your talk proposal

Main Talks 1️⃣ Searching for Shady Patterns: Shining a light on UK corporate ownership - Adam Hill In June 2016 the UK government launched the world’s first “beneficial ownership” register; a requirement for all UK companies to register who were the “persons of significant control”, PSCs, who actually controlled the company. Recent investigative journalism has made headlines with the leaking of the Panama and Paradise papers and it is clear that transparency in corporate ownership needs to be a significant factor within modern democracy. In a partnership between DataKind UK and Global Witness we have built the worlds first network graph mapping all of the UK public data on those who control corporate interests in the UK; it comprises in excess of 4.5 million companies and 4 million individual people. It has been enriched with company officer data and metrics of financial secrecy based upon geographic regions.

The goal of the project was to enable Global Witness to search for "shady patterns" within corporate ownership networks to act as leads for investigative journalism to expose corrupt practices. Further more, we were able to analyse the completeness of the register and identify ways of improving such data structures to inform other world governments how to best build similar public registers of corporate ownership.

We present here how we built this amazing data structure using Python tools for cleaning and data processing and a Neo4j graph database storing the network graph itself. In addition, we share the first insights derived from this process.

2️⃣ Breaking the Black Box - How to Evaluate Your Agents... in Real Time Too! - Craig West If you are building with LLMs, creating high quality evaluations is one of the most impactful things you can do. Without evals, it can be very difficult and time intensive to understand how different model versions might affect your use case. This talk aims to provide you a roadmap that may be simpler than you think to implement.

In this talk, we will look at the two aspects of Observability and Evaluation. Using the manual evaluating-ai-agents.com, along with its code repo, we will see that observability can be done without vendor solutions but with standard Python, either during Evaluation Driven Development or after development.

We will look at three core evaluation strategies - deterministic, human and LLM as Judge - with code examples.

Lightning Talks ⚡ ⚡1️⃣ TBA ⚡2️⃣ TBA

Please note:

1. 🚨🚨🚨A valid photo ID is required by building security. You MUST use your initial/first name and surname on your meetup profile, otherwise, you will NOT make it on the guest list! 🚨🚨🚨
2. This event follows the NumFOCUS Code of Conduct, please familiarise yourself with it before the event.

If your RSVP status says "You're going" you will be able to get in. No further confirmation required. You will NOT need to show your RSVP confirmation when signing in. If you can no longer make it, please unRSVP as soon as you know so we can assign your place to someone on the waiting list.

*** Code of Conduct: This event follows the NumFOCUS Code of Conduct, please familiarise yourself with it before the event. Please get in touch with the organisers with any questions or concerns regarding the Code of Conduct. *** There will be pizza & drinks, generously provided by our host, Carnival UK. ***

Logistics Doors open at 6.30 pm, talks start at 7 pm. For those who wish to continue networking and chatting we will move to a nearby pub/bar for drinks from 9 pm.

Please unRSVP in good time if you realise you can't make it. We're limited by building security on the number of attendees, so please free up your place for your fellow community members!

Follow @pydatasoton (https://twitter.com/pydatasoton) for updates and early announcements. We are also on Instagram/Threads as @pydatasoton, and find us on LinkedIn.

PyData Kampala Meetup – You're Invited! Join us for an exciting in-person Meetup where we explore the power of open source and AI in Africa’s data journey.

Talks: 1. Open Source Data Visualization Opportunities by Kavuma Lameck 2. Building Africa\, One Open AI Tool at a Time by Kaddu Livingstone

Connect with fellow data enthusiasts, share ideas, and discover opportunities shaping Africa’s tech future. Date: Friday 27th June 2025, 5-7pm (Prompt) Venue: Unipod, Yusuf Lule Teaching Facility, Makerere University. Register now: https://forms.gle/43xGkBv4VXMi53ny9

Title: Quantum-centric supercomputing and applications. Abstract: Quantum-centric supercomputing (QCSC) is a new computational framework that makes use of high-performance classical and quantum computing in concert. In this talk, I will review IBM's roadmap for QCSC, illustrating how QCSC algorithms have enabled new use cases in chemistry and materials science, and the role of QCSC in the future of the quantum era. And a short Bio: Javier Robledo-Moreno is a Research Scientist at IBM Quantum. Since he joined IBM two years ago, his research has focused on the development and implementation of quantum-centric supercomputing algorithms to study quantum many-body problems. Javier obtained his PhD in Physics from New York University and the Flatiron Institute in 2023. Hybrid Option: https://teams.live.com/meet/9344585286624?p=VQZ6HlXKEe08HQKtaJ

The Tech Week vibes don't need to end! NYC Women in Machine Learning and Data Science (WiMLDS) is excited to host a summer evening mixer on Wed, June 25th, co-presented with Women in AI, NYAI and Supermomos.

Come ready to meet fellow women in machine learning, data science and AI! All levels welcome.

To join this event, please RSVP on supermomos event page rather than this Meetup invite. Space is limited.

Stay connected: https://www.linkedin.com/company/wimlds-nyc/

Note: Food & drinks are on individual tabs.

Since relaunching in February, we are meeting on every fourth (not last!) Wednesday of the month, and whilst it is the wrong end of the month for that most geeky of celebrations, Star Wars day, it would be remiss not to tip the hat. The venue is the Hand & Racquet in Wimbledon (25-27 Wimbledon Hill Rd, SW19 7NE). https://www.greeneking.co.uk/pubs/greater-london/hand-and-racquet

Come meet like-minded people, with a common interest in Linux. Whether you are a complete neophyte or Linux warrior with decades of experience, come join the chat about distros, networks, open-source software, personal projects, ... or nothing related whatever.

Chris and Gulraj will be sharing hosting duties with at least one of us present at each meetup. We realise that this venue is not as convenient for some people as our central London former haunts, but the pub is only 2 minutes walk from the station, and Wimbledon is well connected (only 17 minutes out of London Waterloo mainline, on the Sutton branch of the Thameslink, and a District line terminus, and the western-most point on the Croydon Tramlink).

The pub is on the flat, at street level, including the toilets, although the space reserved for us (presently, at least) is a raised mezzanine area up a small number of steps at the back of the pub. We want to be inclusive, so if this proves to be problematic we will seek to change arrangements with the pub in the future.

Please come join us. Sign-ups will open shortly before the end of the May gathering.

Directions from the station: Exit the station through the main concourse (near the District Line platforms), turn right along the main road passing the station taxi rank (on Orinoco Lane) as you do so, cross Alexandra Road at the traffic lights. After the gym on the corner, pass the chemist to reach the pub.

🎫Tickets only via Eventbrite 👉 https://www.eventbrite.de/e/1219114978719 Use the promo code " SIBB-PitchNight-2025-MEETUP " to get 25% off! -------------------------------------------------------------------- Once again we are happy to invite you to our annual highlight event: The SIBB Startup Pitch Night. And we are celebrating a premiere: The SIBB Startup Pitch Night event will be part of the great "buildersklub conference". Enjoy high-quality networking with startups, supporters, and investors from Berlin's startup scene, as well as the no-code and low-code community of buildersklub (formerly 'lowcodeklub'). Our Startup Pitch Night will highlight the buildersklub conference evening event: 9 founder teams will enter the stage and present their products and what they've achieved in 10 months SIBB incubator. They have worked hard to turn their ideas into impactful projects and will be thrilled to answer your questions and celebrate this milestone with you.

📅 When: June 25, 2025, 5:30 PM – 11:00 PM 📍 Where: Cambridge Innovation Center Berlin (CIC). Join us for a night of great ideas, inspiring stories, high-value networking, and good vibes!

------------------------------------------------------------ 🎫Tickets only via Eventbrite 👉 https://www.eventbrite.de/e/1219114978719 Use the promo code " SIBB-PitchNight-2025-MEETUP " to get 25% off! ------------------------------------------------------------

PyData Helsinki is Back! 🚀🐍

After a quiet period following five awesome online events in 2020–2021, we're rebooting PyData Helsinki!

With the recent sold-out, standing-room-only Helsinki Python events proving there's plenty of enthusiasm in our community, now is a great opportunity to meet and reconnect in person.

First Meetup:

• 🗓 Wednesday, June 25th
• ⏰ 17:00
• 📍 Kaisla (not Kaivopuisto)

Come along and let's enjoy a relaxed evening of networking and discussions. Meet old friends, make new connections, and let's talk about what we want PyData Helsinki to become!

Please register for the event so we know how many are coming.

We were planning a picnic but the weather forecast doesn't look promising, so let's go to a pub instead. Join the Helsinki Python Discord.

― ― ― ― ― ―

And we've already found a great host for August: Wonna is a hands-on software consulting company that is committed to working with the developer community. We're privileged to be able to have our next meetup there. Save the date:

• 🗓 Tuesday, August 26th
• ⏰ 18:00
• 📍 Wonna, Elimäenkatu 17-19

We need some talks for this event. If you would like to give a talk, please reach out. We welcome talks of all levels from beginner to advanced. Also ⚡️ lightning talks of 5–10 minutes! Suitable subjects include the tools of the data trade, including but not limited to Python, and experiences using them. Think more "this is why Parquet is a great file format" or "I hated Cursor until I implemented these .cursorrules" than "AGI is coming, buy my product or be left behind" or "quantum-proof your digitalisation strategies with AI-driven design sprints".

Meetups always need venues—please ask your employer to host an event!

― ― ― ― ― ―

Looking forward to seeing everyone again, or in many cases for the first time,

The PyData Helsinki Team

We are delighted to announce the 24th London Information Retrieval & AI Meetup, a free evening event aimed at enthusiasts and professionals curious to explore and discuss the latest trends in the field.

This time, the Meetup is Hybrid, with a live event in London streamed online on Zoom!

ATTENTION: Remember to fill out the form to confirm the registration and receive the link to the virtual event: https://forms.gle/UQfBgcCLm2fG7Zza9

>>>> IN-PRESENCE MEETUP

Location:

Gladwin Tower, nine elms point sw8 2fs, London [Google Maps] open doors: 6:15 PM (GMT+1)

>>>> ONLINE MEETUP

Location: Zoom

open doors 6:30 PM (GMT+1) You will receive the virtual event link after completing the form.

----------------------------------------

The event will be structured around 2 technical talks, each followed by a Q&A session. The event will end with a networking session.

> Open doors from 6:15 PM (in-person) > 6:30 PM open doors for virtual attendees

• Welcome & Latest News - Alessandro Benedetti, Director @ Sease

> 06:45 PM First talk Building Search Using OpenSearch: Limitations and Workarounds Nazerke Seidan - IR/ML Engineer @ Sease

> 7:30 PM Second Talk Exploring Multilingual Embeddings for Italian Semantic Search: A Pretrained and Fine-tuned Approach Nicolò Rinaldi - Software Engineer/Data Scientist @ Sease

> 8:15\~8:45 PM Networking session + buffet

For more info, including speaker bios, abstracts and timing, please check our website https://sease.io/

Join us for our QA meetup of the BrowserStack Meetup group Berlin.

Session : Stop Bugs Before They Bite : Shift Left testing by Shweta Varma

Venue: Delivery Hero, Oranienburger Str. 70, 10117 Berlin

Date & Time - June 24th, 6:00 PM CEST

Want to stay connected with the Berlin QA community beyond the meetup? Join our BrowserStack Discord Server to get updates, continue discussions, and network with fellow QA professionals. Don't miss out—be part of the conversation!

Click here to join our BrowserStack Discord Server

Few notes: 🔹 This will be an In-Person event

Mark your calendars and reserve your spot now! We can't wait to see you there!

Note- RSVP doesn't confirm your seat, We will be sending the confirmation email one day prior to the Meetup to confirm your seat.

Join us at our next NYC meetup for an evening of real talk focused on infrastructure scaling, rethinking attention, and cost-efficient pipeline design.

Whether you're scaling foundation models, fine-tuning inference costs, or just trying to make your pipelines less brittle—this one’s for you. Come network with fellow ML engineers, researchers, and infra builders over good food and smarter systems thinking.

🧠 Featuring:

• ​Maxim Alt, Senior Director, AI Software ​ AI Systems Optimizations Maxim Alt is currently working in a Stealth AI Infra Startup. A 30-year veteran of high-performance computing. He founded several tech start-ups in the Bay area and served as SVP of AI & HPC Technology at Core Scientific, after an 18-year engineering career at Intel. Max has authored dozens of research papers on advanced topics in power, performance and efficiency optimization, resource provisioning and prediction, and holds 10 patents for large scale HPC workflows. Max holds a B.Sc. in Math and Computer Science from Tel Aviv University and a M.Sc. in Software Management from Carnegie Mellon University.

• ​Rekha Singhal, Head of Research @ Tata Consultancy Services ​ GenAI from PoC to Production Dr. Rekha Singhal is Head of Research at TCS PacePort New York and leads the Computing Systems Research area at TCS Research. An ACM Senior Member with 70+ patents and 100+ publications, she focuses on accelerating modern and agentic-AI applications across heterogeneous multi-cloud environments. Rekha holds M.Tech. and a Ph.D. in Computer Science from IIT Delhi and was a visiting researcher at Stanford.

• ​Jacob Buckman, Co-founder and CEO @Manifest AI ​ Rethinking Attention Jacob Buckman is Founder & CEO of Manifest AI. They are on a mission to train a neural network to model all human output. Jacob brings years of deep learning research from institutions like OpenAI, GoogleBrain and Mila. Jacob holds a Ph.D in Computer Science from McGill and multiple degrees from Carnegie Mellon University.

Date & Location: 📍 Location: Goodwin Procter LLP 📆 Date: June 23rd, 2025 🕠 Time: 5:30 PM – 8:30 PM ET 🍕 Snacks, drinks, and networking included

Free to attend, space is limited! Reserve your spot now: https://lu.ma/gwt3ikhr

Powered by Banc of California

Expect:

• Deep dives from engineers and researchers
• Actionable insights for ML infra and pipeline design
• A relaxed, high-signal environment with food, drinks, and conversation

PLEASE REGISTER ON LUMA https://lu.ma/gwt3ikhr https://lu.ma/gwt3ikhr https://lu.ma/gwt3ikhr

PyData Pittsburgh is excited to host Jay Palat, from Mpathic, for our first event of the summer. Join us on Thursday, June 19th, as Jay will talk about how machine learning is used to process conversational data.

Billions of conversations happen every day, covering topics from the banal to the mission critical. Many of these conversations are ephemeral and retained only in memories, but more and more are collected in everything from YouTube transcripts to congressional minutes to ‘this call will be recorded for quality assurance’.

Find out how we use Python and machine learning to advance clinical accuracy and quality with AI-powered oversight to enhance patient safety, streamline workflows, and ensure compliance.

About the talk:

This session is for engineers and data scientists interested in extracting info from human conversation. We’ll look at the tools available to process conversations and tradeoffs involved. We’ll talk about the challenges we’ve found in capturing, analyzing and delivering insights for conversations and some of the ways we’ve solved them for our customers' needs. We’ll look at some of the tools and techniques available in the Python ecosystem starting with regular expressions and going all the way to large language models.

Location:

We'll be gathering at COhatch Waterfront, a beautiful coworking space at the Waterfront in Homestead. COhatch coworking space has graciously provided the event space for this Meetup. We know many of our PyData Pittsburgh members work hybrid or remotely. If you are interested in learning more about COhatch coworking spaces and would like a free one day trial, please sign up here.

We want to thank COhatch for investing in our PyData Pittsburgh community!

Times: 6pm, Doors Open 6:30pm - 7:30pm, Panning for Gold in the Streams of Conversation Talk

Subscribe to our Substack. Follow us on Linked In. Interested in giving a Meetup talk? Check out our CFP here.

When June 19\, 2025 \| 10:00 AM Pacific

When and Where Online. Register for the Zoom.

Multi-Modal Rare Events Detection for SAE L2+ to L4

A burst tire on the highway or a fallen motorbiker occur rarely and thus pose extra efforts to Autonomous vehicles. Methods to tackle such edge cases in road scenarios are explained.

About the Speaker

Wolfgang Schulz is Product Owner for Lidar Perception at Continental. He engages in the automotive industry since 2005. With his team he currently works on components for an SAE L4 stack.

Voxel51 + NVIDIA Omniverse: Exploring the Future of Synthetic Data

Join us for a lightning talk on one of the most exciting frontiers in Visual AI: synthetic data. We’ll showcase a sneak peek of the new integration between FiftyOne and NVIDIA Omniverse, featuring fully synthetic downtown scenes of Santa Jose. NVIDIA Omniverse is enabling the generation of ultra-precise synthetic sensor data, including LiDAR, RADAR, and camera feeds, while FiftyOne is making it easy to extract value from these rich datasets. Come see the future of sensor simulation and dataset curation in action, with pixel-perfect labels to match.

About the Speaker

Daniel Gural is a seasoned Machine Learning Engineer at Voxel51 with a strong passion for empowering Data Scientists and ML Engineers to unlock the full potential of their data.

O-TPT: Orthogonality Constraints for Calibrating Test-time Prompt Tuning in Vision-Language Models

We propose O-TPT, a method to improve the calibration of vision-language models (VLMs) during test-time prompt tuning. While prompt tuning improves accuracy, it often leads to overconfident predictions. O-TPT introduces orthogonality constraints on textual features, enhancing feature separation and significantly reducing calibration error across multiple datasets and model backbones.

About the Speaker

Ashshak Sharifdeen is a visiting Student Researcher at Mohamed bin Zayed University of Artificial Intelligence, UAE

Advancing MLLMs for 3D Scene Understanding

Recent advances in Multimodal Large Language Models (MLLMs) have shown impressive reasoning capabilities in 2D image and video understanding. However, these models still face significant challenges in achieving holistic comprehension of complex 3D scenes. In this talk, we present our recent progress toward enabling global 3D scene understanding for MLLMs. We will cover newly developed benchmarks, evaluation protocols, and methods designed to bridge the gap between language and 3D perception.

About the Speaker

Xiongkun Linghu is a research engineer at the Beijing Institute for General Artificial Intelligence (BIGAI). His research focuses on Multimodal Large Language Models and Embodied Artificial Intelligence, with an emphasis on 3D scene understanding and grounded reasoning.

When June 19\, 2025 \| 10:00 AM Pacific

When and Where Online. Register for the Zoom.

Multi-Modal Rare Events Detection for SAE L2+ to L4

A burst tire on the highway or a fallen motorbiker occur rarely and thus pose extra efforts to Autonomous vehicles. Methods to tackle such edge cases in road scenarios are explained.

About the Speaker

Wolfgang Schulz is Product Owner for Lidar Perception at Continental. He engages in the automotive industry since 2005. With his team he currently works on components for an SAE L4 stack.

Voxel51 + NVIDIA Omniverse: Exploring the Future of Synthetic Data

Join us for a lightning talk on one of the most exciting frontiers in Visual AI: synthetic data. We’ll showcase a sneak peek of the new integration between FiftyOne and NVIDIA Omniverse, featuring fully synthetic downtown scenes of Santa Jose. NVIDIA Omniverse is enabling the generation of ultra-precise synthetic sensor data, including LiDAR, RADAR, and camera feeds, while FiftyOne is making it easy to extract value from these rich datasets. Come see the future of sensor simulation and dataset curation in action, with pixel-perfect labels to match.

About the Speaker

Daniel Gural is a seasoned Machine Learning Engineer at Voxel51 with a strong passion for empowering Data Scientists and ML Engineers to unlock the full potential of their data.

O-TPT: Orthogonality Constraints for Calibrating Test-time Prompt Tuning in Vision-Language Models

We propose O-TPT, a method to improve the calibration of vision-language models (VLMs) during test-time prompt tuning. While prompt tuning improves accuracy, it often leads to overconfident predictions. O-TPT introduces orthogonality constraints on textual features, enhancing feature separation and significantly reducing calibration error across multiple datasets and model backbones.

About the Speaker

Ashshak Sharifdeen is a visiting Student Researcher at Mohamed bin Zayed University of Artificial Intelligence, UAE

Advancing MLLMs for 3D Scene Understanding

Recent advances in Multimodal Large Language Models (MLLMs) have shown impressive reasoning capabilities in 2D image and video understanding. However, these models still face significant challenges in achieving holistic comprehension of complex 3D scenes. In this talk, we present our recent progress toward enabling global 3D scene understanding for MLLMs. We will cover newly developed benchmarks, evaluation protocols, and methods designed to bridge the gap between language and 3D perception.

About the Speaker

Xiongkun Linghu is a research engineer at the Beijing Institute for General Artificial Intelligence (BIGAI). His research focuses on Multimodal Large Language Models and Embodied Artificial Intelligence, with an emphasis on 3D scene understanding and grounded reasoning.

AI for the rest of us is on a mission to make AI easier to understand.

AI is ready for everyone. No matter who you are. No matter what your role.

We promise no hype, no jargon and no confusing terminology. We're a warm and inclusive community where we meet you where you are, and help get you where you want to be!

Join us for our next in-person meetup in London on Thursday June 19th

At this event we have two amazing speakers who will provide unique perspectives on the Human benefits and impacts of AI. Azahara Corrales (AI Digital Transformation Leader) will be speaking about how AI can be a tool that empowers people and Rob Price (Founder, Futuria) will be reflecting on Digital Responsibility when delegating work to multi-agent teams.

Where and When?

• Thursday, June 19th
• Doors open at 18:00
• Talks start at 19:00
• Adaptavist, 28 Scrutton St, London EC2A 4RP

Talk 1: AI, Made Human: What You Really Need to Know (Azahara Corrales, Digital Transformation Leader)

In a world where AI can feel distant, overwhelming, or even threatening, this talk brings it back to the human level. Azahara Corrales breaks down what AI really is (and isn’t), and shows how anyone, regardless of background, can start using it today to improve productivity, creativity, and confidence. It's not about replacing humans; it’s about empowering them.

Azahara Corrales is an international public speaker and award-winning AI strategist, known for helping brands and professionals embrace artificial intelligence to drive meaningful transformation. With a background in digital marketing and a passion for closing the tech gap, she empowers organisations and individuals, especially women, to confidently navigate the future of work with AI.

Talk 2: Digital Responsibility with Multi-Team, Multi-Agent Agentic AI (Rob Price, Founder of Futuria)

Rob founded the global Corporate Digital Responsibility movement and brings that knowledge and experience to the domain of multi-agent teams. As we innovate with AI agents we are constantly asking ourselves the questions of "what work do we delegate to teams of AI agents?" and importantly "what does that mean for us?"

Massive thank you to our sponsors DevRev.ai and Kolteki who host our meetups! Without their support this would not be possible!

Code of Conduct This event has a code of conduct that you can review here.

When June 19\, 2025 \| 10:00 AM Pacific

When and Where Online. Register for the Zoom.

Multi-Modal Rare Events Detection for SAE L2+ to L4

A burst tire on the highway or a fallen motorbiker occur rarely and thus pose extra efforts to Autonomous vehicles. Methods to tackle such edge cases in road scenarios are explained.

About the Speaker

Wolfgang Schulz is Product Owner for Lidar Perception at Continental. He engages in the automotive industry since 2005. With his team he currently works on components for an SAE L4 stack.

Voxel51 + NVIDIA Omniverse: Exploring the Future of Synthetic Data

Join us for a lightning talk on one of the most exciting frontiers in Visual AI: synthetic data. We’ll showcase a sneak peek of the new integration between FiftyOne and NVIDIA Omniverse, featuring fully synthetic downtown scenes of Santa Jose. NVIDIA Omniverse is enabling the generation of ultra-precise synthetic sensor data, including LiDAR, RADAR, and camera feeds, while FiftyOne is making it easy to extract value from these rich datasets. Come see the future of sensor simulation and dataset curation in action, with pixel-perfect labels to match.

About the Speaker

Daniel Gural is a seasoned Machine Learning Engineer at Voxel51 with a strong passion for empowering Data Scientists and ML Engineers to unlock the full potential of their data.

O-TPT: Orthogonality Constraints for Calibrating Test-time Prompt Tuning in Vision-Language Models

We propose O-TPT, a method to improve the calibration of vision-language models (VLMs) during test-time prompt tuning. While prompt tuning improves accuracy, it often leads to overconfident predictions. O-TPT introduces orthogonality constraints on textual features, enhancing feature separation and significantly reducing calibration error across multiple datasets and model backbones.

About the Speaker

Ashshak Sharifdeen is a visiting Student Researcher at Mohamed bin Zayed University of Artificial Intelligence, UAE

Advancing MLLMs for 3D Scene Understanding

Recent advances in Multimodal Large Language Models (MLLMs) have shown impressive reasoning capabilities in 2D image and video understanding. However, these models still face significant challenges in achieving holistic comprehension of complex 3D scenes. In this talk, we present our recent progress toward enabling global 3D scene understanding for MLLMs. We will cover newly developed benchmarks, evaluation protocols, and methods designed to bridge the gap between language and 3D perception.

About the Speaker

Xiongkun Linghu is a research engineer at the Beijing Institute for General Artificial Intelligence (BIGAI). His research focuses on Multimodal Large Language Models and Embodied Artificial Intelligence, with an emphasis on 3D scene understanding and grounded reasoning.

When June 19\, 2025 \| 10:00 AM Pacific

When and Where Online. Register for the Zoom.

Multi-Modal Rare Events Detection for SAE L2+ to L4

A burst tire on the highway or a fallen motorbiker occur rarely and thus pose extra efforts to Autonomous vehicles. Methods to tackle such edge cases in road scenarios are explained.

About the Speaker

Wolfgang Schulz is Product Owner for Lidar Perception at Continental. He engages in the automotive industry since 2005. With his team he currently works on components for an SAE L4 stack.

Voxel51 + NVIDIA Omniverse: Exploring the Future of Synthetic Data

Join us for a lightning talk on one of the most exciting frontiers in Visual AI: synthetic data. We’ll showcase a sneak peek of the new integration between FiftyOne and NVIDIA Omniverse, featuring fully synthetic downtown scenes of Santa Jose. NVIDIA Omniverse is enabling the generation of ultra-precise synthetic sensor data, including LiDAR, RADAR, and camera feeds, while FiftyOne is making it easy to extract value from these rich datasets. Come see the future of sensor simulation and dataset curation in action, with pixel-perfect labels to match.

About the Speaker

Daniel Gural is a seasoned Machine Learning Engineer at Voxel51 with a strong passion for empowering Data Scientists and ML Engineers to unlock the full potential of their data.

O-TPT: Orthogonality Constraints for Calibrating Test-time Prompt Tuning in Vision-Language Models

We propose O-TPT, a method to improve the calibration of vision-language models (VLMs) during test-time prompt tuning. While prompt tuning improves accuracy, it often leads to overconfident predictions. O-TPT introduces orthogonality constraints on textual features, enhancing feature separation and significantly reducing calibration error across multiple datasets and model backbones.

About the Speaker

Ashshak Sharifdeen is a visiting Student Researcher at Mohamed bin Zayed University of Artificial Intelligence, UAE

Advancing MLLMs for 3D Scene Understanding

Recent advances in Multimodal Large Language Models (MLLMs) have shown impressive reasoning capabilities in 2D image and video understanding. However, these models still face significant challenges in achieving holistic comprehension of complex 3D scenes. In this talk, we present our recent progress toward enabling global 3D scene understanding for MLLMs. We will cover newly developed benchmarks, evaluation protocols, and methods designed to bridge the gap between language and 3D perception.

About the Speaker

Xiongkun Linghu is a research engineer at the Beijing Institute for General Artificial Intelligence (BIGAI). His research focuses on Multimodal Large Language Models and Embodied Artificial Intelligence, with an emphasis on 3D scene understanding and grounded reasoning.

When June 19\, 2025 \| 10:00 AM Pacific

When and Where Online. Register for the Zoom.

Multi-Modal Rare Events Detection for SAE L2+ to L4

A burst tire on the highway or a fallen motorbiker occur rarely and thus pose extra efforts to Autonomous vehicles. Methods to tackle such edge cases in road scenarios are explained.

About the Speaker

Wolfgang Schulz is Product Owner for Lidar Perception at Continental. He engages in the automotive industry since 2005. With his team he currently works on components for an SAE L4 stack.

Voxel51 + NVIDIA Omniverse: Exploring the Future of Synthetic Data

Join us for a lightning talk on one of the most exciting frontiers in Visual AI: synthetic data. We’ll showcase a sneak peek of the new integration between FiftyOne and NVIDIA Omniverse, featuring fully synthetic downtown scenes of Santa Jose. NVIDIA Omniverse is enabling the generation of ultra-precise synthetic sensor data, including LiDAR, RADAR, and camera feeds, while FiftyOne is making it easy to extract value from these rich datasets. Come see the future of sensor simulation and dataset curation in action, with pixel-perfect labels to match.

About the Speaker

Daniel Gural is a seasoned Machine Learning Engineer at Voxel51 with a strong passion for empowering Data Scientists and ML Engineers to unlock the full potential of their data.

O-TPT: Orthogonality Constraints for Calibrating Test-time Prompt Tuning in Vision-Language Models

We propose O-TPT, a method to improve the calibration of vision-language models (VLMs) during test-time prompt tuning. While prompt tuning improves accuracy, it often leads to overconfident predictions. O-TPT introduces orthogonality constraints on textual features, enhancing feature separation and significantly reducing calibration error across multiple datasets and model backbones.

About the Speaker

Ashshak Sharifdeen is a visiting Student Researcher at Mohamed bin Zayed University of Artificial Intelligence, UAE

Advancing MLLMs for 3D Scene Understanding

Recent advances in Multimodal Large Language Models (MLLMs) have shown impressive reasoning capabilities in 2D image and video understanding. However, these models still face significant challenges in achieving holistic comprehension of complex 3D scenes. In this talk, we present our recent progress toward enabling global 3D scene understanding for MLLMs. We will cover newly developed benchmarks, evaluation protocols, and methods designed to bridge the gap between language and 3D perception.

About the Speaker

Xiongkun Linghu is a research engineer at the Beijing Institute for General Artificial Intelligence (BIGAI). His research focuses on Multimodal Large Language Models and Embodied Artificial Intelligence, with an emphasis on 3D scene understanding and grounded reasoning.