Member of Technical Staff - Code Generation

Who is Recruiting from Scratch

Recruiting from Scratch is a talent firm that focuses on placing the best candidates for our clients. Our team is 100% remote, and we work with companies across North America, South America, and Europe to help them build high‑performing teams.

Location: San Mateo, CA |
Employment Type: Full‑time |
Location Type: On‑site |
Department: Engineering & Research

We’re a frontier AI team building systems that turn natural language into interactive world simulations
. As part of this work, we’re developing autonomous code‑generation agents that can reason, write, execute, and verify code in real environments.

This role is for an Applied AI Research Engineer focused on code generation and agentic systems
, working across post‑training, evaluation, and execution infrastructure.

• Build and improve agentic code‑generation systems (planning, tool use, execution, self‑repair).
• Design control loops (ReAct, Reflexion, Tree‑of‑Thought, Lang Graph‑style).
• Train and post‑train models using SFT, DPO, RLAIF, or RLHF
  .
• Create evaluation suites for multi‑step coding tasks.
• Ship agents that work end‑to‑end on real codebases
  , not just demos.
• Collaborate closely with infrastructure and product teams.

• Experience building or shipping code‑generation agents or AI systems.
• Strong understanding of LLM systems, evaluation, and failure modes.
• Comfort working across research and production code.
• Systems mindset around correctness, safety, and execution.
• Willingness to work on‑site in San Mateo
  .

• Agents that pass real repo test suites end‑to‑end
• Contributions to agent frameworks or OSS evals
• Published work in agentic systems or code generation
• Experience with sandboxed execution or secure environments

• Base salary: $200,000 – $400,000
• Total compensation:
  Up to $1M+, depending on caliber and impact
• Meaningful equity included; compensation is flexible and talent‑first

• Work on frontier AI problems with real‑world impact
• Build production‑grade autonomous agents
• High ownership, fast iteration, and direct influence
• Join early and help shape foundational AI systems