L

Staff ML Engineer, Agent Training & Environments

San Francisco, CA · full time · Lead / Staff / Principal

$250,000 – $280,000

Listed on Labelbox’s own careers site. You apply with them directly — we never stand between you and the employer.

What this role is

Build infrastructure and systems that enable frontier AI labs to train and evaluate agentic models, bridging high-throughput platform engineering with deep reinforcement learning expertise. This role combines designing RL environments and reward systems, implementing verification and grading pipelines, and scaling fine-tuning infrastructure for teams pushing the boundaries of AI agents.

Our summary, not Labelbox’s wording. The full posting is on their site.

Skills this role names

Log in to see which of these are already on your profile.

What they ask for

Required

3+ years shipping production systems others rely on
Strong system and API design judgment
Daily production code shipped with coding agents
Build infrastructure for team tooling, CI, and harnesses
Work effectively in ambiguous startup environments
Deep Python proficiency
Fine-tuned models for agentic tasks using SFT and at least one RL method
Built environments for agents to operate in
Designed verifiers or graders for open-ended work
Forensic debugging of training runs
Understand compute economics and experimental efficiency
Document and communicate learnings to the team

Nice to have

Experience with agent harnesses and coding agents in training and evaluation
Multi-tenancy and sandboxing for untrusted agent execution
Production distributed systems, ML infrastructure, or data systems at scale
Direct experience with frontier labs or highly technical customers