Research Engineer

Role overview

This position is for a Research Engineer concentrating on post-training and reasoning. The role may be full-time or part-time, and the company is open to applicants who are currently doing a PhD or Master’s, as well as professionals already working in industry.

The work centers on advancing post-training optimization and reasoning methods, building new algorithms, and partnering with teams across functions to turn research insights into practical AI systems. You will also work with complex datasets, improve model capability, and contribute to research and development efforts focused on AI performance and interpretability.

What the role involves

• Running research in post-training optimization and reasoning techniques.
• Designing novel algorithms and translating research findings into applied AI use cases.
• Collaborating with cross-functional stakeholders to implement work into advanced systems.
• Studying large and complex datasets to extract useful signals.
• Improving AI models and supporting high-impact R&D initiatives.

Candidate profile

• Solid fundamentals in machine learning engineering.
• Hands-on ability to train and fine-tune large language models end to end using tools such as PyTorch, Hugging Face, vLLM, deepspeed/FSDP, or comparable stacks.
• Practical experience with post-training approaches such as SFT, preference optimization, and RL-based fine-tuning.
• Evidence of research or research-quality engineering through publications, meaningful open-source work, or production ML systems at a lab or frontier company.
• Comfort working with a part-time research lead and independently driving work forward while surfacing trade-offs early.
• Strong interest in applied work with real-world impact, especially in healthcare-related problems.

Core work areas

• Planning and executing SFT, DPO, and RL experiments, including GRPO, PPO, and newer methods, on reasoning traces from an expert network.
• Creating benchmarks and evaluation methods that assess clinical and adjudication reasoning, including the reasoning process rather than only final answers.
• Converting raw expert outputs into high-quality training data through schema design, quality checks, and scaling pipelines.
• Working directly with customers such as frontier labs and healthcare AI companies on custom data and evaluation projects.
• Publishing research when it is beneficial and appropriate.

Why this role may appeal to you

• Work that can influence leading institutional investors and major global organizations.
• Opportunity to operate at the intersection of artificial intelligence, institutional finance, and advanced technology.
• A remote-first setup with flexible hours and a distributed global team.
• Chance to help shape design culture and contribute to building a strong design team during growth.
• Competitive pay together with equity and benefits.

Location and working style

The role can be performed remotely or in person. The company is based in San Francisco, but candidates from anywhere in the world are welcome. Team coordination happens across time zones with shared overlapping core hours.