Database

Wendy Nather is Senior Research Initiatives Director at 1Password and a veteran security leader who previously led advisory CISO programs and research at Duo Security, the Retail ISAC, and 451 Research; she serves on the steering committee of the Institute for Security and Technology’s Ransomware Task Force and on the board of Sightline Security.

No summary available yet.

No summary available yet.

No summary available yet.

Materials Science. I am interested in finding solutions that can unpick the damage in our natural environment.

No summary available yet.

Founding researcher at Geodesic Research leading pretraining and midtraining safety directions to address dangerous capabilities such as biorisk and offensive cyber, and to instill deeper alignment and character into large language models.

3-month stipend to continue working on goal misgeneralisation project for ICLR deadline, plus travel funding

Giuseppe Dal Prá is the Founder and CEO of the Odyssean Institute and a transdisciplinary thinker focused on whole-society transitions to a more resilient and flourishing civilisation. He studied History and Politics at Balliol College, University of Oxford, is a Fellow of the Royal Society of Arts, and has written on preparing democratic institutions for wicked problems and systemic risks such as runaway climate change using deliberative and other antifragile reforms. He also co-authored the ComplexCity proposal, a semi-finalist in the Global Challenges Foundation’s New Shape Prize, which outlined an epistemic democratic platform for citizen-led policy experimentation.

Nathan Lile is a co-founder and CEO of SynthLabs.ai, a research lab focused on improving the reasoning and alignment of frontier AI systems using reinforcement learning from human and AI feedback and synthetic data. He previously worked as a researcher at Stability AI and continues to collaborate with EleutherAI on post-training methods such as RLHF and RLAIF for GPT-NeoX and related large language model projects.([scholar.google.com](https://scholar.google.com/citations?hl=en&user=y6v6c-4AAAAJ))

No summary available yet.

No summary available yet.

Anja Manuel is a co-founder and principal at Rice, Hadley, Gates & Manuel LLC, a strategic consulting firm that helps U.S. companies navigate international markets and geopolitical risk. She is also executive director of the Aspen Strategy Group and Aspen Security Forum and previously served in the U.S. Department of State as special assistant to the under secretary for political affairs, focusing on South Asia policy.

Computing resources and researcher salaries at a new deep learning + AI alignment research group at Cambridge

No summary available yet.

No summary available yet.

No summary available yet.

No summary available yet.

No summary available yet.

Wyatt Tessari L'Allié is the Founder and Executive Director of AI Governance and Safety Canada (AIGS Canada), a non-partisan not-for-profit organization based in Ottawa working to ensure advanced AI is safe and beneficial. He moved to Ottawa in October 2021 to build the Canadian AI safety and governance community, consolidating various Canadian AI safety groups into a unified national organization. His background spans mechanical engineering (École Polytechnique de Montréal, 2011), filmmaking (New York Film Academy, 2008), mining and emissions compliance work, and climate advocacy through three electoral campaigns between 2010 and 2015. After learning about AGI and x-risk in 2015, he spent four years researching 21st-century challenges and authored a book on global issue prioritization before transitioning fully to AI governance work. Since 2022, AIGS Canada has made public policy submissions to the Canadian federal government and he has testified before multiple parliamentary committees on AI governance. He received Long-Term Future Fund grants to connect, expand, and enable the AGI safety and governance community in Canada.

No summary available yet.

Software engineer at Elicit

4-month stipend for upskilling within the field of economic governance of AI

No summary available yet.

Tahir Mahmood Saleh is passionate about building a better, more inclusive world. With hands-on experience in project management, governance, climate action, and youth empowerment, he has led and supported impactful initiatives that uplift communities and drive change. Through his work with organizations like Bridge Nigeria Initiative, DOTAKEACTION, OCCEN, and now with Center for Renewable Energy and Action on Climate Change CREACC-NG, Tahir champions citizen engagement, economic empowerment, and sustainable development, always bringing energy, empathy, and a vision for a brighter future.

Jonathan Ng is a Singapore-based AI safety researcher and engineer currently working as a Project Officer at the Singapore AI Safety Institute (AISI). He completed the SERI MATS 3.1 program (Spring 2023), where he worked with the Cadenza Labs team on extending the 'Discovering Latent Knowledge' paper by contributing to the EleutherAI/elk library, running large-scale hyperparameter sweeps, and conducting original experiments in LLM probing. He holds a BComp in Computer and Information Systems Security from the National University of Singapore (2017-2021). His research includes co-authoring the MACHIAVELLI benchmark (ICML 2023), which measures trade-offs between rewards and ethical behavior in AI agents, and the Catastrophic Cyber Capabilities Benchmark (3CB, AAAI 2025 Workshop). He is also a co-author of CCS-Lib, a Python package for eliciting latent knowledge from LLMs published in the Journal of Open Source Software. Beyond research, Jonathan has been a key organizer of AI safety capacity-building in Singapore, founding the Research Engineering Camp for Alignment Practitioners (RECAP), directing AI Safety Fellowships at NUS and NTU, and serving as an instructor at the ML4Good Singapore bootcamp. He received funding from the Long-Term Future Fund for his early ML alignment skill development and his SERI MATS project.

5-month part time salary for collaborating on a research paper analyzing the implications of compute access

Research Manager at MATS, London

Nicole Ruiz is a non-resident fellow at the Foundation for American Innovation and director of The Third Oikos, a publication exploring how technology reshapes family life and the household. She is also a visiting fellow at the Institute for Family Studies and previously invested in frontier technology startups at seed stage as a venture capitalist at Compound.

Subhojeet Pramanik is a researcher at Softmax with an MSc from Alberta’s RLAI lab whose work focuses on reinforcement learning and representation learning. He introduced the Recurrent Linear Transformer (AGaLiTe, TMLR 2024), previously worked at IBM Research, and is the author of OmniNet, a unified architecture for multi-modal multi-task learning.

Ondrej Bajgar is a DPhil student in Bayesian machine learning at the University of Oxford, supervised by Michael A. Osborne, Alessandro Abate, and Konstantinos Gatsis. He studied mathematics at the University of Warwick and subsequently spent three years as a Research Scientist at IBM Watson, working on machine learning for text understanding and dialogue systems. He then joined the Future of Humanity Institute at Oxford as a Senior Research Scholar, where he worked on AI safety for approximately two years before beginning his doctorate. His current research focuses on active inverse reinforcement learning, developing methods to align AI decision-making with human preferences by strategically selecting informative demonstrations. He has also published work on negative human rights as a principled framework for long-term AI safety and regulation, co-authored with Jan Hořeňovský in the Journal of Artificial Intelligence Research. He received LTFF funding to support his AI safety PhD at Oxford.

No summary available yet.

MATS 3-month stipend to use singular learning theory to explain & control the development of values in ML systems

Technical AI risk analyst at ORCG and systems engineer from Universidad de los Andes, co‑founder of AI Safety Colombia, focusing on advanced AI safety and mechanistic interpretability.

Looking for Fund for my PhD at Oxford - Interested in AI Alignment, Safety and Governance.

Purchasing a technologically adequate laptop for my AI Policy studies in the ML Safety Scholars program and at Oxford

Senior Marketer at Probably Good with a master’s degree in Communication Science and ongoing master’s studies in Philosophy; serves as a Marketing Advisor for Effectief Geven and has previously worked for Amnesty International, the United Nations, and advised think tanks, government bodies, and tech companies on communication strategy and impact.

6-month salary for independent alignment research in interpretability or control

Research agenda aimed at developing methods for constructing powerful, easily interpretable world-models.

No summary available yet.

No summary available yet.

Niko McCarty is a molecular bioengineer, writer, and Fellow at Astera Institute whose work focuses on biophysics, metascience, and visual tools for explaining complex ideas. He previously worked as a data journalist, founded and serves as founding editor of the biotechnology magazine Asimov Press, and is writing a book on cell biology titled “Biology is a Burrito & Other Essays.”

No summary available yet.

No summary available yet.

No summary available yet.

No summary available yet.

Virtual AI Safety Unconference 2024 is a collaborative online event by and for researchers of AI safety

Glauber De Bona is a Brazilian computer scientist and assistant professor in the Department of Computer Engineering and Digital Systems (PCS) at the Escola Politécnica, University of São Paulo (USP). He holds a Bachelor's in Computer Engineering from the Aeronautics Institute of Technology (ITA, 2006), a Master's in Computer Science from USP (2011), and a PhD in Computer Science from USP (2016), with his doctoral dissertation focused on measuring inconsistency in probabilistic knowledge bases. He conducted postdoctoral research at University College London (2017) and USP (2018). His academic research spans probabilistic logic, inconsistency measurement and localization in knowledge bases, epistemic argumentation, and formal epistemology, with publications in venues such as AAAI and the Journal of Artificial Intelligence Research. He has described his broader research interest as mitigating existential risks posed by AI, and has been active on the Alignment Forum under the handle glauberdebona, where he has posted work on computational complexity arguments about the alignment problem. He received a 6-month stipend from the Long-Term Future Fund to transition to independent research on AI safety.

Amritanshu Prasad is an AI safety researcher and policy writer based in the United Kingdom, focused on AI evaluations, AI governance, and the intersection of nuclear risk with emerging AI. He received an upskilling grant to study the AGI Safety Fundamentals Alignment Curriculum, and has since built a career in technical AI safety work, including a research engineering role at Equistamp where he helped UK AISI develop ControlArena, an infrastructure project for evaluating protocols for controlling unaligned frontier LLMs. He has contracted with METR for baselining and model interaction tasks, and facilitated courses on AI governance and transformative AI through BlueDot Impact. He is a member of the working group on international AI governance at the Alva Myrdal Centre for Nuclear Disarmament at Uppsala University, co-authoring work on the influence of nuclear governance frameworks on AI governance. He also completed a research fellowship at Pivotal Research and is founding Suav Tech, a for-profit AI safety evaluations organization. He writes the Substack newsletter The Next Frontier, covering AI evals, AI policy, and nuclear risk.