Bay Area Frontier Research Club #17: Founders Edition | AGI House SF (dinner + research discussions)
This is our first Founders Edition session: two technical founders present the open research problems inside their companies, paper-discussion style — real data, falsifiable hypotheses, and deep Q&A.
One talk on trust — verifying AI-generated content the moment it's made, not after the damage is done. One talk on physics — teaching foundation models how factories actually behave, from raw and unruly sensor data. Two founders who've bet their companies on the answers.
The Frontier Research Club is a curated forum for rigorous, technical discussion at the frontier of AI. We convene researchers from the frontier labs, Stanford, Berkeley, and the teams building in production to examine concrete work — papers, methods, and results — with a bias toward assumptions, evaluation methodology, failure modes, and what would count as convincing evidence.
Each session features 2–3 talks selected for rigor and discussion value. Presentations are intentionally brief so the majority of time is reserved for questions and critique. Papers and supporting materials are shared in advance to ensure a high-baseline conversation.
Agenda
5:30pm: Doors open
5:30pm – 6:30pm: Networking + light dinner
6:30pm – 8:00pm: Research presentations + discussion
8:00pm – 8:30pm: Networking
Presenters & topics
TALK 1: Does Sensor Data Need Its Own Foundation Model?
Linda Hong Cheng — Founder, Abstract Atomic
Foundation models have transformed language, vision, video, and increasingly robotics. But much of the physical world is observed through a very different modality: heterogeneous sensor telemetry, sparse operational events, control inputs, physical constraints, and delayed outcomes. Using manufacturing and battery sensor data as the test case, Linda asks what a foundation model for physical dynamics should actually look like — examining time-series foundation models, graph-based dynamics, physics-informed learning, and hybrid architectures that learn freely but verify against physical constraints.
Linda Hong Cheng is the founder of Abstract Atomic, building higher-order foundation models for manufacturing — starting with battery and industrial production systems. A Clarendon Scholar at the University of Oxford (Nuffield College), her research spans physical-world foundation models and computational social science, including work published in the Oxford Handbook of the Sociology of Machine Learning. Abstract Atomic won the Oxford Catalyst Grant, and Linda has spoken on AI at Google, IBM, and Davos-week stages.
TALK 2: Verification, Not Detection: Research Directions in Real-Time Trust Layers for AI-Generated Media
Ankit Shah — Founder
As AI-generated voice, video, and text scale, the ability to distinguish what is real, who is acting, and whether intent is legitimate continues to erode. Detection-first approaches remain structurally reactive. This talk examines a verification-first research orientation: checks oriented toward claims compliance, platform policy, and brand safety prior to distribution — focusing on the underlying open problems of multimodal verification under real-world latency and deployment constraints, and scalable learning from imprecise labels at internet scale, drawing on the NeurIPS 2024 framework for imprecise label learning.
Ankit Shah is LLM Architecture Associate Director at Accenture's Center for Advanced AI. He holds a PhD in Language Technologies from Carnegie Mellon University (MLSP group, advised by Bhiksha Raj and Rita Singh), with publications spanning NeurIPS, ICLR, and ICASSP on imprecise-label learning, LLM unlearning, and multimodal machine learning. His research interests include systems for verifying AI-generated content. Previously: deep learning at ReviveMed, verification engineering at ARM. He contributed to Fortune Analytics with Fortune Media (iF Design Award 2025; related patent application assigned to Accenture).
Want to present your work?
If you have a research paper you’d like to discuss at one of our next sessions, please submit it for consideration.
Submit your paper here!
Who should attend
Technical founders and founding engineers
Researchers in multimodal ML, trust & safety, time-series/world models, and physics-informed learning
ML engineers working on media authenticity, industrial AI, or sensor data
Investors backing AI infrastructure and physical-world AI
Capacity is limited.
We will take photos and short video clips for event recap and promotion. By attending, you consent to being photographed and recorded, and to the use of those images and clips by the organizers on social media and other event marketing channels.
🌐 Connect with Frontier Research Club
Luma Calendar: luma.com/frontiersyndicate
YouTube: FrontierResearchClub
LinkedIn: frontier-research-club
Instagram: @frontierresearchclub
Email: kristopher@frontiersyndicate.vc
Hosted by
Frontier Syndicate is a venture community connecting frontier tech researchers, builders, and investors through curated convenings and early-stage capital. Across the Bay Area, we host a recurring series of research forums, builder nights, and intimate investor dinners — and back exceptional companies emerging from the labs, communities, and technical networks we convene.
AGI House SF is a community of AI founders and researchers accelerating humanity's transition to AGI, hosting merit-based gatherings, hackathons, and technical events that draw leading AI minds from around the world.
Made possible by
Levangie Laboratories is an AI research and development company building causal learning infrastructure for autonomous agents. Their Causal Trajectory Learning architecture preserves the causal chain between an agent's context, decisions, outcomes, and feedback — so agents can retrieve relevant experience, improve across sessions, and make that improvement inspectable without retraining the underlying model.
They're hiring engineers and researchers.