SlipstreamJobs tracks this role from the company's public career site. Apply directly on the employer's site.
Salary: USD 245,000 - 300,000 / annual
Pareto.AI builds the platform that converts expert human judgment into training data, evaluations, and reinforcement learning environments for frontier AI models. You'll own the end-to-end design, development, and production operation of RL environments that power post-training loops at leading labs like Anthropic and GDM.
In this role, you'll partner closely with research teams to translate rough training goals into buildable specifications, architect and implement containerized RL environments and MCP tools, write graders that score task performance, and maintain production health across deployed systems. You'll sit at the intersection of Pareto's engineering team and frontier lab researchers, responsible for catching misalignment between research intent and technical feasibility early.
Key responsibilities include: scoping problems with research requesters, building production-grade Docker images and tools, designing evaluation harnesses, shipping environments into customer platforms, investigating failed tasks and degraded pipelines, and feeding platform gaps back into the product roadmap. You'll invest in build and release infrastructure to reduce delivery time and cost for subsequent environments, moving beyond hand-rolled one-offs.
You bring 7+ years of production systems experience with strong Python or TypeScript skills and proven expertise building and shipping containerized services. You understand Docker layering, dependency pinning, and reproducibility at scale. You're fluent with AI coding agents and can critically review their output. You've owned production systems through incidents, debugged your own cloud deployments on AWS or equivalent platforms, and written documentation that prevents recurrence. You're comfortable with evolving specs, high-contact stakeholder engagement, and ownership through the full lifecycle including post-launch production support. You're based in the US and can travel to the Bay Area as needed.
