talk-data.com

IAQF & Thalesians Seminar Series: Building Trustworthy Agentic AI for Finance: From Knowledge Discovery to Governed Execution. How agentic systems turn financial intent into auditable action—beyond prompts, beyond models.

A Seminar by Stefano Pasquali

6:00 PM Seminar Begins

7:30 PM Reception

Hybrid Event

Fordham University McNally Amphitheater 140 West 62nd Street New York, NY 10023

Free Registration! For Virtual Attendees: Please email [email protected] for the link.

Abstract: Agentic AI in finance is shifting from conversational assistance to systems that execute actions across data, models, and workflows. This evolution raises a central challenge: how to build AI that is not only capable, but trustworthy in high-stakes, regulated environments.

This talk presents a practical framework for building trustworthy agentic AI for finance, structured around four building blocks: knowledge discovery on unstructured data, knowledge discovery on structured data, multi-agent reasoning and orchestration, and governance with continuous interpretability and surveillance. We argue that trust is a system-level property, emerging from how these components interact rather than from model accuracy alone.

As a focused deep dive, we examine Text-to-SQL as a stress test for trustworthy execution, highlighting challenges such as schema grounding, semantic precision, and material correctness. Using recent financial benchmarks as reference, we show how agent orchestration and governance layers transform promising capabilities into production-ready financial systems.

Bio: Stefano Pasquali is Senior Vice President and Head of Financial Services Solutions at Domyn, where he leads the creation of sovereign, explainable, and fully auditable AI for mission-critical financial applications. His work combines agentic AI architectures, knowledge-graph reasoning, proprietary financial models, and large-scale LLMs, enabling institutions to own, govern, and scale AI with full compliance and IP protection.

Previously Managing Director and Head of Investment AI at BlackRock, he advanced AI, machine learning, and GenAI to support investors and the Aladdin platform. Earlier he directed Liquidity Research at Bloomberg, where he contributed to BVAL and created Liquidity Assessment (LQA). A theoretical and computational physicist, he has taught at Rutgers and Columbia University and remains active in global conferences and regulatory dialogues on responsible AI and model governance.

Welcome to Women in Data Science - New York City May 2024!

WiDS New York City is independently organized by Chuxin Liu, Kannu Priya, Maria Sanchez, Niki Karanikola, Paulina Toro Isaza, Ruchi Mahindru, Sandhya Nayak, Sneha Varghese, Srideepika Jayaraman, Thara Palanivel, Vanishri Murali, Vasanthi Gopal to be part of the mission to increase participation of women in data science and to feature outstanding women doing outstanding work.

Our aim is to inspire, educate and support Women in the field of Data Science, by uplifting, energizing and building long term relationships across companies, partners, communities and universities in the New York and Tri-State region.

Join us at CUNY Graduate Center for a day filled with inspiring talks, workshops, and networking opportunities. Whether you're a seasoned data scientist or just starting out, this event is perfect for anyone interested in the world of data science. Learn from industry experts, connect with like-minded individuals, and gain valuable insights into the latest trends in data science. Don't miss this chance to be part of a supportive community of women in the data science field. Mark your calendars and get ready for an unforgettable experience at Women in Data Science - New York City!

All genders are invited to attend all WiDS Worldwide & Regional conference events.

Agenda:

11:15 AM to 12:00 PM - Resume building workshop (Optional)

12:00 PM - 1:00 PM - Networking - All are welcome to join! (Bring your own lunch)

1:00 PM - 1:05 PM - Welcome note and introduction to WIDS

1:05 PM - 1:35 PM - Keynote speech: "Future of Work: A Vision for AI Powered Digital Labor" by Dr. Gargi Dasgupta, Director, Technical Product Management and AI Leader, Digital Labor @ IBM

1:35 PM - 2:00 PM - "AI for Women's Health" by Dr. Yiye Zhang, Associate Professor, Division of Health Informatics, Department of Population Health Sciences, Graduate Faculty, Cornell Systems Engineering, Informatics Director, Clinical Decision Support

2:00 PM - 2:25 PM - "Red Teaming Generative Language Models" by Dr. Ioana Baldini, Senior Research Scientist, IBM Research AI Organization

2:25 PM - 2:50 PM - Break

2:50 PM - 3:00 PM - Recognize Fact Vs Fiction - AI Edition by WiDS Ambassadors Srideepika Jayaraman & Vanishri Murali

3:00 PM - 3:25 PM - "Generative AI Use Cases in the Public Sector" by Aakanksha Joshi, Senior AI Engineer @IBM

3:25 PM - 3:50 PM - "Seeding, Weeding and Feeding - Keys to Successful Marketing” by Madhavi Chakrabarty, Sr. Director and Asst. Professor of Professional Practice, Rutgers Business School

3:50 PM - 4:35 PM Panel on Career Development Panel Moderators - Chuxin Liu - Senior Associate @ J.P. Morgan Chase Maria Sanchez - Senior Client Engineering AI Engineer @IBM Panelists: Niki Karanikola Machine Learning Engineer @VNA Eda Kavlakoglu - Program director @IBM Amaris Efthimiou - Innovation Designer @ IBM Client Engineering Ruchi Mahindru - IBM Distinguished Engineer

4:35 PM - 4:40 PM - Update on PyData Community Events

4:40 PM - 4:55 PM - Lucky draw event - Attendees (both Online and Live participants) qualify to win at the lucky draw to get 30 mins coaching/mentoring sessions with our esteemed Speakers & WiDS Ambassadors. Make sure to register and mark your attendance.

4:55 AM - 5:00 PM - Vote of Thanks

Gareth Ford Williams is a pioneer of digital accessibility. His substantial contributions to the BBC and the wider field have helped shape a more inclusive digital landscape for people with disabilities. His breadth of expertise, spanning standards, fonts, inclusive technologies, and advocacy speaks volumes about the positive impact he has had on countless lives.

Zoom Link

https://voxel51.com/computer-vision-events/september-14-meetup/

ARMBench: An Object-Centric Benchmark Dataset for Robotic Manipulation

Amazon Robotic Manipulation Benchmark (ARMBench), is a large-scale, object-centric benchmark dataset for robotic manipulation in the context of a warehouse. ARMBench contains images, videos, and metadata that corresponds to 235K+ pickand-place activities on 190K+ unique objects. The data is captured at different stages of manipulation, i.e., pre-pick, during transfer, and after placement from a robotic workcell in an Amazon warehouse. Benchmark tasks are proposed by virtue of high-quality annotations and baseline performance evaluation are presented on three visual perception challenges, namely 1) object segmentation in clutter, 2) object identification, and 3) defect detection. ARMBench can be accessed at http://armbench.com.

Chaitanya Mitash is Sr. Applied Scientist at Amazon Robotics. His research focuses on computer vision and manipulation for item manipulation. He received his Ph.D. in computer science from Rutgers University.

Fan Wang is a Research Scientist at Amazon Robotics, with a focus on robotic manipulation and perception. She holds a Ph.D. in electrical and computer engineering from Duke University. Her undergraduate degree was in electrical and mechanical engineering from the University of Edinburgh, UK.

Mani Nambi is a Sr. Applied Scientist at Amazon Robotics. His research focuses on manipulation systems for item and package handling. He received his Ph.D. in mechanical engineering from the University of Utah.

From Model to the Edge, Putting Your Model into Production

This talk delves into the journey from model training to deployment at the edge – an often neglected yet vital aspect of machine learning implementation. It elucidates the essential practices and challenges associated with transitioning an AI model from a controlled environment to real-world edge devices.

Joy is a Machine learning Engineer at Secury360, a startup offers a hardware box that turns your security cameras into a perimeter security system with no false detections. He is responsible for the model training, active learning infrastructure and managing the labeling team. If you have been in the FiftyOne Slack you probably have seen him around.

Optimizing Distributed Fine-Tuning Workloads for Stable Diffusion with the Intel Extension for PyTorch on AWS

In this talk, we explore the use of the Intel Extension for PyTorch to optimize a vision generative AI workload. The vision workload focuses on the fine-tuning of a stable diffusion model, on the AWS cloud using Intel’s 4th Generation Xeon Processors. We leverage optimizations like Intel Advanced Matrix Extensions (AMX) and mixed-precision with BF16 and FP32, to speed up training. Attendees can expect (1) A technical dive into the workload and solution (2) a brief code walkthrough (3) workload setup on AWS, and (4) a short demo.

Eduardo Alvarez is a Senior AI Solutions Engineer at Intel and a specialist in applied deep learning and AI solution design. His background includes building software tools for the energy sector, and his primary interests lie in time-series analysis, computer vision, and cloud solutions architecture. Additionally, he is a community leader in data science and ML/AI for geosciences.

Zoom Link

https://voxel51.com/computer-vision-events/september-14-meetup/

ARMBench: An Object-Centric Benchmark Dataset for Robotic Manipulation

Amazon Robotic Manipulation Benchmark (ARMBench), is a large-scale, object-centric benchmark dataset for robotic manipulation in the context of a warehouse. ARMBench contains images, videos, and metadata that corresponds to 235K+ pickand-place activities on 190K+ unique objects. The data is captured at different stages of manipulation, i.e., pre-pick, during transfer, and after placement from a robotic workcell in an Amazon warehouse. Benchmark tasks are proposed by virtue of high-quality annotations and baseline performance evaluation are presented on three visual perception challenges, namely 1) object segmentation in clutter, 2) object identification, and 3) defect detection. ARMBench can be accessed at http://armbench.com.

Chaitanya Mitash is Sr. Applied Scientist at Amazon Robotics. His research focuses on computer vision and manipulation for item manipulation. He received his Ph.D. in computer science from Rutgers University.

Fan Wang is a Research Scientist at Amazon Robotics, with a focus on robotic manipulation and perception. She holds a Ph.D. in electrical and computer engineering from Duke University. Her undergraduate degree was in electrical and mechanical engineering from the University of Edinburgh, UK.

Mani Nambi is a Sr. Applied Scientist at Amazon Robotics. His research focuses on manipulation systems for item and package handling. He received his Ph.D. in mechanical engineering from the University of Utah.

From Model to the Edge, Putting Your Model into Production

This talk delves into the journey from model training to deployment at the edge – an often neglected yet vital aspect of machine learning implementation. It elucidates the essential practices and challenges associated with transitioning an AI model from a controlled environment to real-world edge devices.

Joy is a Machine learning Engineer at Secury360, a startup offers a hardware box that turns your security cameras into a perimeter security system with no false detections. He is responsible for the model training, active learning infrastructure and managing the labeling team. If you have been in the FiftyOne Slack you probably have seen him around.

Optimizing Distributed Fine-Tuning Workloads for Stable Diffusion with the Intel Extension for PyTorch on AWS

In this talk, we explore the use of the Intel Extension for PyTorch to optimize a vision generative AI workload. The vision workload focuses on the fine-tuning of a stable diffusion model, on the AWS cloud using Intel’s 4th Generation Xeon Processors. We leverage optimizations like Intel Advanced Matrix Extensions (AMX) and mixed-precision with BF16 and FP32, to speed up training. Attendees can expect (1) A technical dive into the workload and solution (2) a brief code walkthrough (3) workload setup on AWS, and (4) a short demo.

Eduardo Alvarez is a Senior AI Solutions Engineer at Intel and a specialist in applied deep learning and AI solution design. His background includes building software tools for the energy sector, and his primary interests lie in time-series analysis, computer vision, and cloud solutions architecture. Additionally, he is a community leader in data science and ML/AI for geosciences.

Zoom Link

https://voxel51.com/computer-vision-events/september-14-meetup/

ARMBench: An Object-Centric Benchmark Dataset for Robotic Manipulation

Amazon Robotic Manipulation Benchmark (ARMBench), is a large-scale, object-centric benchmark dataset for robotic manipulation in the context of a warehouse. ARMBench contains images, videos, and metadata that corresponds to 235K+ pickand-place activities on 190K+ unique objects. The data is captured at different stages of manipulation, i.e., pre-pick, during transfer, and after placement from a robotic workcell in an Amazon warehouse. Benchmark tasks are proposed by virtue of high-quality annotations and baseline performance evaluation are presented on three visual perception challenges, namely 1) object segmentation in clutter, 2) object identification, and 3) defect detection. ARMBench can be accessed at http://armbench.com.

Chaitanya Mitash is Sr. Applied Scientist at Amazon Robotics. His research focuses on computer vision and manipulation for item manipulation. He received his Ph.D. in computer science from Rutgers University.

Fan Wang is a Research Scientist at Amazon Robotics, with a focus on robotic manipulation and perception. She holds a Ph.D. in electrical and computer engineering from Duke University. Her undergraduate degree was in electrical and mechanical engineering from the University of Edinburgh, UK.

Mani Nambi is a Sr. Applied Scientist at Amazon Robotics. His research focuses on manipulation systems for item and package handling. He received his Ph.D. in mechanical engineering from the University of Utah.

From Model to the Edge, Putting Your Model into Production

This talk delves into the journey from model training to deployment at the edge – an often neglected yet vital aspect of machine learning implementation. It elucidates the essential practices and challenges associated with transitioning an AI model from a controlled environment to real-world edge devices.

Joy is a Machine learning Engineer at Secury360, a startup offers a hardware box that turns your security cameras into a perimeter security system with no false detections. He is responsible for the model training, active learning infrastructure and managing the labeling team. If you have been in the FiftyOne Slack you probably have seen him around.

Optimizing Distributed Fine-Tuning Workloads for Stable Diffusion with the Intel Extension for PyTorch on AWS

In this talk, we explore the use of the Intel Extension for PyTorch to optimize a vision generative AI workload. The vision workload focuses on the fine-tuning of a stable diffusion model, on the AWS cloud using Intel’s 4th Generation Xeon Processors. We leverage optimizations like Intel Advanced Matrix Extensions (AMX) and mixed-precision with BF16 and FP32, to speed up training. Attendees can expect (1) A technical dive into the workload and solution (2) a brief code walkthrough (3) workload setup on AWS, and (4) a short demo.

Eduardo Alvarez is a Senior AI Solutions Engineer at Intel and a specialist in applied deep learning and AI solution design. His background includes building software tools for the energy sector, and his primary interests lie in time-series analysis, computer vision, and cloud solutions architecture. Additionally, he is a community leader in data science and ML/AI for geosciences.

Zoom Link

https://voxel51.com/computer-vision-events/september-14-meetup/

ARMBench: An Object-Centric Benchmark Dataset for Robotic Manipulation

Amazon Robotic Manipulation Benchmark (ARMBench), is a large-scale, object-centric benchmark dataset for robotic manipulation in the context of a warehouse. ARMBench contains images, videos, and metadata that corresponds to 235K+ pickand-place activities on 190K+ unique objects. The data is captured at different stages of manipulation, i.e., pre-pick, during transfer, and after placement from a robotic workcell in an Amazon warehouse. Benchmark tasks are proposed by virtue of high-quality annotations and baseline performance evaluation are presented on three visual perception challenges, namely 1) object segmentation in clutter, 2) object identification, and 3) defect detection. ARMBench can be accessed at http://armbench.com.

Chaitanya Mitash is Sr. Applied Scientist at Amazon Robotics. His research focuses on computer vision and manipulation for item manipulation. He received his Ph.D. in computer science from Rutgers University.

Fan Wang is a Research Scientist at Amazon Robotics, with a focus on robotic manipulation and perception. She holds a Ph.D. in electrical and computer engineering from Duke University. Her undergraduate degree was in electrical and mechanical engineering from the University of Edinburgh, UK.

Mani Nambi is a Sr. Applied Scientist at Amazon Robotics. His research focuses on manipulation systems for item and package handling. He received his Ph.D. in mechanical engineering from the University of Utah.

From Model to the Edge, Putting Your Model into Production

This talk delves into the journey from model training to deployment at the edge – an often neglected yet vital aspect of machine learning implementation. It elucidates the essential practices and challenges associated with transitioning an AI model from a controlled environment to real-world edge devices.

Joy is a Machine learning Engineer at Secury360, a startup offers a hardware box that turns your security cameras into a perimeter security system with no false detections. He is responsible for the model training, active learning infrastructure and managing the labeling team. If you have been in the FiftyOne Slack you probably have seen him around.

Optimizing Distributed Fine-Tuning Workloads for Stable Diffusion with the Intel Extension for PyTorch on AWS

In this talk, we explore the use of the Intel Extension for PyTorch to optimize a vision generative AI workload. The vision workload focuses on the fine-tuning of a stable diffusion model, on the AWS cloud using Intel’s 4th Generation Xeon Processors. We leverage optimizations like Intel Advanced Matrix Extensions (AMX) and mixed-precision with BF16 and FP32, to speed up training. Attendees can expect (1) A technical dive into the workload and solution (2) a brief code walkthrough (3) workload setup on AWS, and (4) a short demo.

Eduardo Alvarez is a Senior AI Solutions Engineer at Intel and a specialist in applied deep learning and AI solution design. His background includes building software tools for the energy sector, and his primary interests lie in time-series analysis, computer vision, and cloud solutions architecture. Additionally, he is a community leader in data science and ML/AI for geosciences.