Research Engineer

About this role

Tessera Labs seeks a Research Engineer to build the post-training and evaluation infrastructure for a multi-agent AI platform that transforms enterprise systems. You'll own the training, environment, and inference machinery that turns research hypotheses into production models, working on reinforcement learning for long-horizon agent behavior in complex, verifiable enterprise landscapes.

What you'll do

• Design and scale post-training stacks including SFT, preference optimization, and reinforcement learning for tool use and enterprise transformation tasks
• Build memory and context systems for long-horizon agents, including retrieval, compaction, and training mechanisms
• Develop representation layers using ontologies and knowledge graphs derived from real enterprise systems
• Create data generation and curation pipelines including synthetic landscapes, transformation traces, and curriculum infrastructure
• Build RL environments with sandboxed execution and verification harnesses for automated checking and scoring
• Implement offline evaluation infrastructure for agentic behavior with trajectory-level scoring and reproducible task suites

What they're looking for

• Reinforcement learning systems and long-horizon agent training
• Post-training optimization (SFT, preference optimization, RLHF)
• Python and systems-level programming
• Data pipeline design and curation at scale
• Knowledge graphs and ontology representation
• Training infrastructure optimization and distributed systems
• Evaluation methodology for complex agentic tasks
• Enterprise systems and domain modeling