Database

Veterinary Student building AI evaluation frameworks in food systems and animal health across Africa

No summary available yet.

Assistant Professor of Computer Science at the University of Virginia who leads the Responsible AI for Science and Engineering (RAISE) group, focusing on foundational challenges in artificial intelligence, privacy, safety and the intersection between machine learning and optimization.

No summary available yet.

No summary available yet.

3 months relocation from Chad to London to work on Eliciting Latent Knowledge with Jake Mendel from Apollo Research

Andrei Alexandru is an AI safety researcher and machine learning engineer currently working at iGent AI on Maestro, an autonomous agentic coding agent. He holds an MPhil in Machine Learning from the University of Cambridge, where his dissertation examined the inductive biases of shallow neural networks, funded by a Long-Term Future Fund grant. He previously worked on the dangerous capability evaluations team at OpenAI and at Atla, where he was first author on the Atla Selene Mini paper introducing a state-of-the-art small language model-as-a-judge. Earlier in his career he participated in the SERI MATS program under Evan Hubinger's mentorship and the ML for Alignment Bootcamp at Redwood Research. His research interests include mechanistic interpretability, AI evaluations, and understanding deceptive alignment, and he maintains the blog inwaves.io where he writes about AI safety topics.

Curriculum Developer & Instructor at the Center for Applied Rationality. Jack studied psychology and gender and has previously worked in roles including event planning, data research, and legal assistance, and is especially interested in gaming, structured analysis, and deep conversations about culture and identity.

Peter Barnett is a researcher at the Machine Intelligence Research Institute (MIRI), where he works on the Technical Governance Team focusing on international coordination and governance strategies to prevent catastrophic outcomes from advanced AI. He holds a Master's degree in Physics from the University of Otago in New Zealand, where he conducted research in quantum optics and quantum fluids. He transitioned into AI safety work through the first cohort of the MATS (ML Alignment Theory Scholars) program and subsequently worked at the Center for Human-Compatible AI on reward learning before joining MIRI in December 2022, initially on technical alignment and later shifting to technical governance. His governance research includes verification mechanisms for international AI agreements, distributed training oversight, and AI capability red-lines. He has authored a book on AI x-risk aimed at policymakers and the general public, and is active on LessWrong and the Alignment Forum under the handle peterbarnett.

Nicholas Kees Dupuis is an AI alignment researcher and entrepreneur who co-founded Mosaic Labs, a nonprofit R&D organization developing AI-facilitated group deliberation tools, where he serves as CEO alongside co-founder and CTO Sofia Vanhanen. He holds a Master's degree in Artificial Intelligence from the University of Groningen (2022), where his thesis focused on theory of mind for multi-agent coordination. He conducted research at EleutherAI under Kyle McDonell and Laria Reynolds, studying how finetuning methods can cause alignment failures analogous to those in reinforcement learning, and has published academic work on computational social choice including the paper "Condorcet Markets" (2023). He is best known in the AI safety community for his influential "Cyborgism" essay on the Alignment Forum (February 2023), which proposed a strategy of using human-in-the-loop systems to safely accelerate alignment research by empowering human agency rather than outsourcing it; he subsequently led the Cyborgism track at AI Safety Camp 2023. He received a $120,000 one-year grant from the Long-Term Future Fund to continue developing his research agenda on making LLMs directly useful for alignment research without advancing capabilities. His personal website is lovedoesnotscale.com and he maintains a Substack under the same name.

Co-founder and co-director of Geodesic Research, leading technical AI safety work within the Meridian Cambridge community.

Connor Dunlop is a researcher and strategist working at the intersection of AI governance, compute security, and geopolitics, with experience spanning frontier AI policy research, hardware‑enabled verification, foresight, and institutional design.

Journalism executive and philanthropy leader who serves as chair of the Hudson Institute Board of Trustees and previously ran business operations at Commentary magazine; she also sits on the board of the Flourishing Future Foundation, which supports research on controlling advanced AI to ensure a safe and flourishing future.

No summary available yet.

No summary available yet.

Building high signal AI community infrastructure in Los Angeles focused on talent, collaboration, and ecosystem development.

I've been invited as I have contributed with input to prepare the meeting.

Nick Corvino is a Tarbell AI Journalism Fellow and writer for ChinaTalk. He holds an M.A. from the Yenching Academy of Peking University, where he studied Chinese philosophy and religion, and a B.A. in philosophy and international relations from Northwestern University.

No summary available yet.

James Lester leads the Oxford AI Safety Initiative and has written about AI safety topics on the Effective Altruism Forum.

No summary available yet.

Geoffrey Miller is an evolutionary psychologist and associate professor of psychology at the University of New Mexico. He is known for his research on sexual selection and human nature and for books such as The Mating Mind and Spent, and in recent years he has become an active public commentator on AI risk and other existential threats.

No summary available yet.

Funding to hire a junior researcher on a 3-5 month contract to help launch projects tackling existential risk

No summary available yet.

6-month salary to skill up and gain experience to start working on AI safety full-time

Travis Moore is the Founder and Executive Director of TechCongress, a nonpartisan nonprofit that recruits, trains, and places technologists in Congress through the Congressional Innovation Fellowship and related programs. He previously spent six years on Capitol Hill as Legislative Director to Rep. Henry A. Waxman, former chair of the House Energy and Commerce Committee, where he focused on technology policy. Moore launched initiatives including Congress’s first digital communications training program, its first Congressional staff conference, and the Congressional Digital Service Fellowship, and he co‑founded #CongressToo, a network of former Congressional staffers that helped bring the #MeToo movement to Capitol Hill. He holds a BS in Marketing from Miami University (OH) and a master’s degree in Contemporary European Politics from the University of Bath.

No summary available yet.

No summary available yet.

No summary available yet.

Kai Sandbrink is a DPhil candidate in computational cognitive neuroscience at the University of Oxford's Department of Experimental Psychology, based at Lady Margaret Hall. He is co-supervised by Professor Christopher Summerfield at Oxford and Professor Wulfram Gerstner at EPFL, where he is also an invited guest researcher. Prior to Oxford, he completed an MS in Neural Systems and Computation at ETH Zurich and an MA in China Studies at Peking University. His research uses deep reinforcement learning as a task-driven model of human behavior, with a focus on learning dynamics, cognitive flexibility, and exploration-exploitation trade-offs. His AI-safety-relevant work includes improving deep learning's understanding of uncertainty and designing safer, more interpretable reward functions for reinforcement learning algorithms. He is an affiliate at Concordia AI and has an interest in East-West cooperation on AI safety and governance. He received a Long-Term Future Fund grant in 2021 for starting funds and moving costs related to his DPhil project.

Funding to hire an in-house entrepreneur on a 12 month contract to investigate a space within existential risk and then found an impactful project

Second-year AI student at Inland Norway University of Applied Sciences, currently on exchange at UC Berkeley. I build ML systems and research-agent tools focused on deployment optimization, LLM serving, reproducibility, and reliable autonomous experimentation.

No summary available yet.

No summary available yet.

James Payor is an independent AI alignment researcher based in Australia. He worked as Research Staff at the Machine Intelligence Research Institute (MIRI) in Berkeley from January 2018 to April 2022, and has since been pursuing independent AI alignment research. Before MIRI, he was co-founder and CTO of Draftable, a document comparison software company based in Melbourne, from 2015 to 2017. His research focuses on agent foundations, proof-based cooperation, corrigibility, and building AI systems that maintain robust alignment with human input. He is notable for developing a method for proof-based cooperation that does not require Löb's theorem, and has published work on modal fixpoint cooperation on the AI Alignment Forum. More recently he has been working on better foundations for theorem proving, computer-assisted mathematics, and dependently-typed programming languages. In his youth he represented Australia at the 2013 International Olympiad in Informatics, earning a silver medal.

No summary available yet.

David Duvenaud is an Associate Professor at the University of Toronto and Canada CIFAR AI Chair whose research spans machine learning, AI safety and AI governance. He co-directs the Schwartz Reisman Institute for Technology and Society, co-founded the Vector Institute for Artificial Intelligence, and previously led the Alignment Evaluations team at Anthropic. He also serves as a director of The AI Safety Foundation and has advised AI companies such as Cohere.

Building a CPU-native AI architecture where meaning is geometric position, not statistical prediction. 27 experiments completed, 5 papers in draft.

No summary available yet.

No summary available yet.

Pooja Tope-Puranik is an AI researcher and machine learning engineer working primarily in healthcare, specializing in early breast cancer detection with AI; she has held ML roles at organizations such as Omdena and AI Directions and serves as a UAE ambassador for women-in-tech communities, promoting women’s participation in STEM.

4 months of stipend for MATS extension work in London studying the safety implications of LLM self-recognition

No summary available yet.

No summary available yet.

Independent research and upskilling for one year, to transition from academic philosophy to AI alignment research

Fund unique approaches to research, field diversification, and scouting of novel ideas by experienced researchers supported by PIBBSS research team

No summary available yet.

Early-stage work on a small internal-control layer that tracks instability in LLM reasoning and switches between SAFE / WARN / BREAK modes.