DataPhilly Tech Talks: Building Adaptive and Agentic AI Systems

Join us at BrainDo for February’s DataPhilly Tech Talks, featuring two talks that explore how modern AI systems learn from human signals and act autonomously in real-world environments. This evening connects cutting-edge research with hands-on engineering, highlighting the evolution from intelligent perception to agentic execution.
Our Host and Sponsor, BrainDo is a Philadelphia-based team of digital marketing experts who partner with clients to create meaningful experiences with data-driven results. They host and manage Philly Analytics, including the monthly Web Analytics Wednesdays community meetups.
🔗 phillyanalytics.org | Web Analytics Wednesday LinkedIn Group

🗓️ Agenda
6:00 – 6:30 PM: Arrival, networking, and refreshments
6:30 – 7:10 PM: Talk 1 – Tony Siu, followed by Q&A
7:10 – 7:50 PM: Talk 2 – Ben Morss, followed by Q&A
7:50 – 8:30 PM: Networking

Tony Siu: Generalizable Detection of Student Engagement in Online Learning Environments - Direct Preference Optimization.
Abstract: Automated recognition of student engagement in online learning is crucial as it enables teachers to adapt content delivery to improve learning. In this paper, we explore a method that finetunes a pretrained vision language model (VLM) to recognize student engagement markers in still images. Our model learns to avoid incorrect answers during finetuning by using the emerging direct preference optimization techniques on self-generated preference pairs based on the correct and incorrect VLM answers. On publicly available student engagement datasets, our model shows superior performance over other approaches and substantially better generalizability over the traditional vision methods. My docs, CAIP published paper
Bio: Tony Siu is an AI Engineer and Multi-modal generative AI researcher with deep experience spanning applied machine learning, computer vision, and large-scale data systems. He has led end-to-end development of state-of-the-art GenAI products, including natural-language-to-game generation systems and vision-based agentic interfaces, while driving major performance and cost optimizations in production environments.
Tony has conducted published research across top-tier venues such as IEEE, CAIP, and ECAI, with a focus on multi-modal preference optimization, engagement modeling, and visual question answering. His background includes research leadership with Google, engineering roles across startups and enterprise teams, and hands-on work with CUDA-accelerated vision systems. He is also an active community builder, founding Code&Coffee Philadelphia to connect engineers, founders, and researchers across the tech ecosystem.

Ben Morss https://benmorss.com: MCP revealed: How does MCP transform an LLM into an agent? Let's learn how MCP works, how to use it, and how to build a server and an app.
Abstract: Anthropic’s Model Context Protocol (MCP) lets your favorite LLM do things in the real world - send emails, make Jira tickets, and browse the web. It lets your LLM be an agent.
We’ll explain how MCP works in detail. We’ll show you how to set up and use a server. We’ll build a little server of our very own. And we’ll take a quick look at the very latest thing - MCP Apps and ChatGPT Apps!
Bio: As DeepL’s Developer Evangelist, Ben works to help anyone access DeepL’s world-class AI experiences and language translations. Previously, at Google, he was a Product Manager on Chrome and a Developer Advocate for a better web. Before that he was a software engineer at the New York Times and AOL, and once he was a full-time musician. He earned a BA in Computer Science at Harvard and a PhD in Music at the University of California at Davis. You might still find him making music with the band Ancient Babies, analyzing pop songs at Rock Theory, and writing a musical that’s not really about Steve Jobs.

👉 RSVP now to secure your spot and join us for an evening of cutting-edge AI insights, great conversations, and community connections!
Please note, by RSVPing this event you agree to our Code of Conduct