Lead research on AI capabilities and behaviors using our AI Village platform.
Design and execute analyses and experiments to discover important findings.
Work independently to drive research efforts and contribute to scaling the Village.
We build interactive AI demos and explainers to help make sense of the future. We're a team of four, focused on long-term open-ended AI agent research.
Investigate and evaluate early-stage grant opportunities focused on transformative AI safety and governance.
Build partnerships and manage relationships with applicants and grantees to support high-impact projects.
Contribute to improving grantmaking processes and tooling, including LLM-assisted workflows.
The Centre for Effective Altruism (CEA) stewards the effective altruism movement, applying evidence, reason, and compassion to solve global poverty, animal suffering, and existential risks. The organization has grown from 42 to 66 core staff, with a culture of ambitious growth and high impact.
Set and evolve the research direction for A1’s core intelligence, including context representation, memory, reasoning, planning, and orchestration.
Define evaluation frameworks that measure real-world usefulness, robustness, safety, and long-term behavior.
Own alignment, safety, and guardrail strategy as first-class product concerns.
A1 builds a proactive smart assistant for everyday users to bring intelligence to conversations, errands, organizing, and workflows. We are a small, high-talent-density team focused on shipping high-quality work and learning at rapid speed.
Conduct red-team evaluations to identify jailbreaks, prompt injections, and misuse scenarios in conversational AI models.
Develop creative adversarial prompts and scenarios to systematically probe model behavior and uncover weaknesses.
Generate high-quality human evaluation data by annotating failures and classifying vulnerabilities.
The partner company is a technology organization focused on AI safety and responsible AI development. They work with a remote, asynchronous team to improve the robustness of conversational AI systems.
Define and own a portfolio of FROs advancing AI resilience, from defensive biosecurity to compute governance.
Source ideas, build founding teams, and coach founders on technical roadmaps and fundraising.
Serve on multiple FRO boards and represent Convergent externally with funders and partners.
Convergent Research is a nonprofit science studio that builds Focused Research Organizations (FROs) to solve technological bottlenecks. They have built over 14 high-impact scientific organizations spanning AI, biotech, and climate science, fostering a collaborative and mission-driven culture.
Apply deep subject-matter expertise to AI model evaluation and large language model projects.
Develop challenging domain-specific problems and assess AI responses for accuracy and reasoning.
Collaborate with AI research teams to improve training datasets and evaluation methodologies.
Jobgether is an AI-powered job matching platform that connects candidates with hiring companies. It offers a remote, asynchronous work culture and uses AI tools to support recruitment.
Own full-cycle recruitment for AI Research, ML, data, evaluations, and ML infra roles.
Partner with research leadership and the CEO to define highly specialised candidate profiles.
Source talent from AI labs, research organisations, universities, and open-source communities.
White Circle is an AI Safety company building the safety, reliability, and optimization layer for AI systems. We are a small, highly focused team of under 50 people backed by top investors from OpenAI, Anthropic, and DeepMind.
Design and build the AI execution platform with event-triggered workflows and model-agnostic runtimes.
Develop evaluation layers with golden test suites and safety checks to ensure AI reliability.
Implement governance mechanisms and build AI agents supporting business workflows.
Jobgether is a platform that connects talent with opportunities using AI-powered matching. The company has a globally distributed team and a remote-first culture.
Conduct adversarial testing on conversational AI to identify vulnerabilities.
Generate human evaluation data and document findings reproducibly.
Communicate risks to both technical and non-technical stakeholders.
AI safety evaluation company that tests and improves conversational AI systems for vulnerabilities and biases. Operates with a remote-first culture and collaborates on diverse projects to promote responsible AI development.
Own the shared AI foundation used by product engineering teams, including model selection, routing, context management, and evaluation.
Design and ship agentic capabilities end to end, from proof-of-concept through production optimization and iterative improvement.
Establish evaluation infrastructure and practices to measure model and agent performance, ensuring quality and accuracy for financial decision-making.
Our partner is building a suite of financial planning and analysis capabilities powered by a shared AI foundation. They operate as a remote-first engineering team with a focus on autonomy, craftsmanship, and high-impact work.
Conduct structured horizon-scanning for novel AI-enabled hazard classes lacking historical base rates.
Investigate next-generation detection sciences like adaptive pathogen surveillance for emerging hazards.
Pursue feasibility analyses on deep resilience and lifeline capacities for engineered infrastructure failures.
CARMA works to help society navigate the complex and potentially catastrophic risks from advanced AI systems, focusing on lowering risks to humanity and the biosphere. It is a fiscally-sponsored project of Social & Environmental Entrepreneurs, Inc., a small nonprofit public benefit corporation with a collaborative culture.
Build prototypes supporting AI verification and turn emerging ideas into practical tools.
Support technical communications and help translate prototypes into policy impact.
Collaborate with external experts across ML engineering, cybersecurity, and confidential computing.
Singapore AI Safety Hub is an international collaboration building open tools to verify AI datacenter activity for global governance and trustworthy AI adoption. It is a young, fast-moving organization with a small core team and partners including Oxford and the Future of Life Institute.
Design and run experiments to understand model behavior and test new ideas.
Create datasets, benchmarks, and evaluation methods for hard problems.
Work directly with AI labs to turn open-ended goals into concrete research projects.
Vetto builds the infrastructure for next-generation AI training data. They partner with the world’s top AI labs and offer a fast, flat, remote-first culture with high ownership.
Architect and build production AI-agent systems like AIDA and NOVA, shaping how Yuno applies LLMs and agentic AI to payment problems.
Design conversational and multilingual agents that hold real customer conversations across channels, optimized for reliability and low latency.
Define evaluation, safety, and guardrails for agentic systems in a compliance-sensitive, multi-market environment.
Yuno is the AI-native operating system of global commerce, powering financial infrastructure for enterprise merchants, banks, and wallets. It connects over 1,000 payment methods across 190+ countries, trusted by global brands like McDonald's and Rappi, with a remote-first culture.
Conduct hands-on adversarial testing across AI models, applications, and data pipelines to identify vulnerabilities.
Perform advanced red-team assessments including jailbreaks, guardrail bypass, and prompt injection analysis.
Produce detailed vulnerability reports and collaborate with AI safety and engineering teams to improve security.
This company specializes in AI security research and adversarial machine learning. It is a remote-first organization with a collaborative culture focused on technical impact and professional growth.
Design and ship AI agent systems end to end, owning data access, retrieval, and integration with existing infrastructure.
Build evaluation frameworks from scratch, including golden datasets and regression harnesses, and think adversarially about prompt injection and data boundary violations.
Set technical direction for AI at Kota and collaborate across teams to surface new AI opportunities.
Kota is reimagining insurance and retirement benefits for the modern workforce. We have raised over €20M, serve tens of thousands of employees, and operate with a remote-first culture.
Assess AI-enabled systems against regulatory requirements, standards, and conformity criteria.
Support clients and internal teams with AI/ML expertise and practical guidance.
Develop methodologies, tools, and training for AI conformity assessment.
This company focuses on the assessment of AI-enabled products and software across regulated industries, including medical devices and robotics. They operate in a global, multidisciplinary environment with a culture centered on trust, safety, and responsible AI adoption.
Coordinate with external experts and contractors across various domains to ensure quality contributions.
Support data projects by helping secure and license data, ensuring compliance.
Organize events such as symposiums and own external communication channels.
Epoch AI is a research institute investigating trends in machine learning and the economic consequences of AI. They have a small, collaborative team with a startup-like culture committed to building an inclusive and supportive community.
Lead research and engineering to develop autonomous AI agents that set and execute complex goals.
Design and implement dynamic planning, memory, tool-use, and evaluation frameworks for safe agent behavior.
Collaborate with multidisciplinary teams to transition research into scalable, production-ready solutions.
Our partner is a large-scale healthcare environment focused on advancing autonomous AI systems. They are looking for a Lead Data Scientist to shape next-generation AI agents in a collaborative, remote U.S. team.
Build sandboxed environments that wrap real Niural workflows, stateful across episodes and seeded for reproducibility.
Design programmatic verifiers from known ground truth, scoring trajectories to prevent reward hacking.
Train agents via RL loops, curriculum schedules, and fine-tuning, then publish negative results internally.
Niural is a global Payroll, Employer of Record (EOR), Agent of Record (AOR), and Contractor Management platform that empowers businesses in the digital economy. We are a team building foundational internet infrastructure with a focus on speed, ownership, and ambition.