Engineering
AI Infrastructure / Product Engineer
Remote · Full-time
We are looking for an AI Infrastructure / Product Engineer to help turn AI research and experiments into reliable, production-ready products at Irys. You will work on AI benchmarking infrastructure, productize research ideas into product features, and build the full-stack and cloud systems that ship them. This is a good fit for someone who wants to work close to AI research while still building real production systems. The work is hands-on, practical, and product-focused. Not demo theater. Real systems, real users, real reliability problems.
Responsibilities
Build benchmarking systems for our AI models, agents, and workflowsSet up evaluation pipelines that measure accuracy, reliability, hallucinations, latency, cost, and retrieval quality across repeatable testsBuild dashboards and reports that track AI system performance over timeTurn AI research ideas and prototypes into usable product features, including RAG systems, legal document workflows, AI agents, drafting tools, and reasoning pipelinesMove experiments from notebooks and scripts into stable production systems, and build the internal tools that help the team test, deploy, and improve AI features fasterBuild backend services, APIs, and product-facing features, with frontend support when neededDeploy and maintain AI systems on cloud infrastructure, including databases, monitoring, CI/CD, and production reliability
Requirements
Strong Python skillsExperience with cloud deploymentFull-stack experience preferredComfortable working with APIs, databases, backend systems, and deployment pipelinesInterest in LLMs, RAG, AI agents, evaluation systems, and applied AI productsAbility to work with ambiguous research ideas and turn them into clean, usable systems
Nice to have
Experience with FastAPI, Docker, React/Next.js, Postgres, vector databases, or cloud platformsExperience building AI evaluation pipelines or benchmarking systemsPrior work with legal-tech, document AI, enterprise AI, or workflow automation