Research Engineer (AI)

Location: Hybrid, based in San Francisco, United States

Company

hirify.global works on frontier AI research and production-impacting projects with research labs and AI companies.

What you will do

• Identify relevant papers, benchmarks, blog posts, and prior work for research directions, then implement or reimplement applicable methods.
• Build reproducible internal research processes and identify improvements to existing methods.
• Own research projects end to end, including scoping, MVP implementation, validation, and delivery of RL or agentic environments and multimodal benchmarks.
• Partner with strategic project and technical leads to turn ambiguous requirements into concrete, testable research plans.
• Validate ideas through implementation, annotation, evaluation, and data sourcing.
• Produce customer datasets, pilots, internal datasets, papers, blog posts, or conference presentations, while assessing technical feasibility for new opportunities.

Requirements

• MS or PhD in ML, CS, or a related quantitative field, or equivalent demonstrated research experience through publications, significant open-source research work, or industry research.
• Strong machine learning research depth, including understanding how models are trained and evaluated, reading papers, assessing their claims, and reimplementing methods.
• Hands-on experience with at least one of reinforcement learning or agentic systems, AI/ML evaluation and benchmarking, or multimodal ML.
• Strong Python skills and the engineering ability to build and ship experiments, evaluation harnesses, environments, and infrastructure independently.
• Ability to turn ambiguous directions into concrete research plans and identify issues proactively.
• Clear technical writing skills.

Nice to have

• Publication track record, preferably as a first author.
• Experience with agent or multimodal benchmarks such as OSWorld, MMMU, WebArena, SWE-bench, or similar, or experience building RL environments or gyms.
• Familiarity with reward modeling, reward hacking, or verifier and judge reliability.
• Familiarity with synthetic data generation or human-in-the-loop workflows.
• Experience with cloud infrastructure and containerized environments.
• Deep reinforcement learning background.

Culture & Benefits

• Full-time hybrid work based in San Francisco.
• Work alongside a multidisciplinary, multinational team.
• Collaborate with frontier labs and prominent AI companies on projects affecting model performance and agent capabilities.
• Opportunities to publish research and present at conferences.

Hiring process

• Only shortlisted candidates will be contacted for an interview.
• Equal opportunity hiring.