Research Engineer, Infrastructure
Core
Build the infrastructure that powers post-training (SFT and RL) for open-weight models used in legal work, including distributed training systems, agent environments, and internal tooling.
Role type
Senior IC research engineer (ML infrastructure)
Builds
Distributed training and RL rollout systems, reproducible agent environments, and internal data management/evaluation tools for legal AI models.
Domain
Artificial Intelligence / Machine Learning / Legal Tech
Deliverable
production ML models
Required skills
Python, distributed systems, PyTorch/JAX, debugging across stack, performance optimization, reliability engineering, observability, data versioning, model serving.
Preferred skills
RL rollouts, agent environments, sandboxed tool execution, post-training workflows.
Technologies
PyTorch, JAX, Python, GPU clusters, distributed training frameworks.
Responsibilities
Scale post-training stack for SFT and RL; build distributed rollout infrastructure for model inference and tool execution; create reproducible agent environments for training and evaluation; own performance and reliability of training runs; identify and remove bottlenecks in training and inference; build internal tools for data management, evals, and job tracking; integrate models into production for serving and online evaluation.
Seniority
Senior, hands-on IC
