Top AI Research Jobs in San Francisco Bay Area, CA

Lead AI evaluation research by defining quality criteria, designing human and automated studies, building rubrics, datasets, benchmarks, and thresholds, and analyzing model performance. Validate automated graders and LLM-as-judge methods, identify failure patterns and uncertainty, and translate findings into model, product, and release decisions. Collaborate with AI researchers, engineers, and cross-functional stakeholders while developing reusable evaluation frameworks and mentoring peers.

Conduct frontier machine learning research to improve Claude’s engineering and scientific capabilities. Design reinforcement learning environments, rewards, training data, benchmarks, evaluation frameworks, simulators, and verifiers. Research long-horizon reasoning, multi-agent collaboration, real-world feedback, robustness, calibration, and generalization. Scale successful approaches from prototypes into production training runs while collaborating with domain experts, industry partners, researchers, and engineers.

Develop and train machine learning models for cloud microphysics, storm evolution, and weather-seeding impacts. Build data pipelines using balloon, radar, satellite, and model data; lead preprocessing, training, evaluation, and uncertainty experiments; and translate research into production forecasting tools. Collaborate with atmospheric and hardware teams on field integration, applying spatiotemporal, physics-informed, or time-series methods and distributed computing.

Conduct technical AI safety and alignment research by designing control protocols, monitoring systems, evaluations, threat models, tools, and datasets for frontier AI systems. Investigate deception, sabotage, reward hacking, multi-agent risks, monitorability, and recursive self-improvement. Build ML pipelines and evaluation harnesses, run experiments, collaborate with industry, government, and academia, and publish findings and policy-oriented technical reports.

Build evaluation frameworks, datasets, sandbox environments, and metrics for production AI agents. Design experiments to compare models, tools, harnesses, and execution environments; investigate failures; and improve agent reliability, performance, and cost. Contribute to agent infrastructure, reusable knowledge systems, internal tooling, and integrations while collaborating with research, engineering, infrastructure, security, and business teams. This applied research role combines production-quality software engineering with open-ended experimentation.

Conduct independent and collaborative research on flexible-data large language models, including Mixture-of-Experts, long-context modeling, retrieval, training dynamics, scaling laws, and data curation. Lead high-impact research projects, train and release models, develop open-source tools, publish scientific papers, and mentor junior researchers. The role is a postdoctoral appointment with opportunities for academic mentorship, grant writing, and teaching.

Lead embedded assessments of frontier AI labs’ safety practices. Scope evaluations, audit existing experiments, identify coverage gaps, design safety experiments, lead teams of 3–5 evaluators, and own the scientific content of public reports. Between engagements, conduct practice evaluations, train scientific staff, and improve evaluation tooling with engineers. The role requires in-person work at host labs and includes collaboration with AI developers, governments, and civil society organizations.

Develops methods to improve foundation models for drug discovery through targeted data generation, model evaluations, and post-training recipes. The role collaborates with pre-training researchers, research engineers, and biological scientists to address protein, antibody, and small-molecule modeling problems. Candidates should bring strong quantitative and machine learning expertise, software engineering depth, and familiarity with biological modeling methods such as AlphaFold, RFDiffusion, ProteinMPNN, and ESM.

The Head of Research will build and lead Amigo’s clinical AI research function, establish evaluation standards, oversee benchmarks and patient simulations, and design studies demonstrating safe, clinically valid agent behavior. The role includes hiring researchers, managing research programs, publishing findings, developing academic collaborations, engaging clinicians and regulators, communicating system limitations, and translating research results into product and deployment decisions.

Lead research on large language models and agentic systems from hypothesis through production deployment. Design post-training experiments using SFT, RLHF, RLAIF, DPO, and GRPO; develop evaluation benchmarks, synthetic data pipelines, and adversarial datasets; shape model architecture and training roadmaps; publish at leading venues; and mentor researchers and engineers.