Bay Area Frontier Research Club | Berkeley Skydeck (dinner + discussion)
The Bay Area Frontier Research Club is a curated forum for rigorous discussion on how AI is reshaping the scientific research process. We convene experimental researchers, computational scientists, and research engineers across domains to examine concrete work—papers, methods, and workflows—covering literature synthesis, hypothesis generation, experimental design, simulation, analysis, and reproducibility.
For each session, we curate 2–3 papers selected for rigor and discussion value. Presentations are intentionally brief so the majority of time is reserved for questions and critique: assumptions, evaluation methodology, failure modes, and what would constitute convincing evidence. Papers and supporting materials are shared in advance to ensure a high-baseline conversation.
Agenda
5:30pm: Doors open
5:30pm – 6:30pm: Networking + snacks
6:30pm – 8:00pm: Research presentations + discussion
8:00pm – 8:30pm: Networking
Presenters & topics
Talk #1: Danny Sawyer (Research Scientist, Google DeepMind)
Exploring Exploration with Foundation Agents in Interactive Environments
Foundation models have been evaluated mostly in static, single-turn settings — but interactive agents need to actively discover information in dynamic environments, not just react to what's given. Danny will present new work investigating how foundation models explore in controlled, zero-shot settings compared to known optimal policies, spanning text-based environments and embodied simulation. The research connects directly to one of the core open questions in agentic AI: can foundation agents learn to explore systematically rather than relying on undirected trial and error?
Review the paper here
Presented by Danny Sawyer - a Research Scientist at Google DeepMind investigating agentic exploration in text and embodied environments. He worked on the SIMA 2 project, which debuted the first multi-domain instructable agent for open-world video games. Previously, he built biologically inspired vision systems for robots deployed in production at Vicarious. He holds a Ph.D. in Bioengineering from Caltech.
Talk #2 — Vignesh Baskaran (Co-Founder, Emily)
Self-Improving Agents: From Human-Designed to Open-Ended Evolution
What happens when you let agents redesign themselves? Vignesh will walk through two recent papers — Darwin Gödel Machine (DGM) and Group-Evolving Agents (GEA) — that demonstrate how individual coding agents can recursively evolve their own architectures and tools, using empirical benchmarks to validate each self-modification. The work points toward a shift from human-designed agent pipelines to open-ended, self-improving systems that autonomously refine how they operate — with significant performance gains on complex coding tasks.
Review the first paper here
Review the second paper here
Presented by Vignesh Baskaran - CTO and Co-Founder of Hexo AI, where the team is building Emily — an autonomous AI scientist for hard, open-ended scientific problems. He holds an Advanced Master's in AI from KU Leuven, ranks in the top 1% on Kaggle, and recently presented at the Precision Medicine World Conference (PMWC 2026) alongside speakers from Stanford, UCSF, and Yale.
Talk #3 — Xinyu Zhang (Member of Special Project, Anyscale)
Ray: A Distributed Framework for Emerging AI Applications
As AI workloads shift from static inference to continuous, interactive systems — agents that learn, explore, and adapt in real time — the underlying infrastructure has to keep up. Xinyu will present the Ray framework, a distributed system designed to unify task-parallel and actor-based computation under a single execution engine, scaling beyond 1.8 million tasks per second. The talk connects directly to the infrastructure challenge behind everything else on tonight's agenda: how do you build systems that can actually run the next generation of agentic AI at scale?
Presented by Xinyu Zhang - Member of Special Projects at Anyscale, the company behind the Ray distributed computing framework. Previously a Software Engineer at Meta, where he applied LLMs to Facebook Groups at scale reaching 100M+ users. He studied at UC Berkeley, where he led the JIPCAD 3D modeling research program.
Want to present your work?
If you have a research paper you’d like to discuss with a cross-disciplinary room, submit it for consideration. Submissions will be considered for future events if not selected this time.
Who should attend
Experimental researchers
Computational scientists across domains (bio/chem/materials/climate/neuro/physics)
Research engineers + lab automation people
Folks building tools for literature review, experiment planning, robotics, simulation, or scientific data
No ML background required. If you’ve ever wished research moved faster, you belong here.
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.
Hosted by
Emily (“MLE”) is an AI scientist that generates novel hypotheses, autonomously designs and runs experiments, critiques its own results, and iterates toward real progress on hard, unsolved problems. Built to augment researchers with agentic, closed-loop experimentation, Emily helps scientists reach the next frontier faster.
Frontier Syndicate is a private venture circle for frontier tech builders, researchers, and investors. We convene high-trust rooms and back exceptional companies at the frontier.
Berkeley SkyDeck is UC Berkeley’s startup accelerator and global innovation hub, supporting founders and researchers bringing breakthrough technologies into the world. SkyDeck helps connect the Berkeley ecosystem—faculty, labs, students, and founders—around ambitious ideas and the people building them.