Senior Machine Learning Engineer

Role overview Own machine learning engineering work from early prototypes through production deployment, with a focus on language model systems for tasks such as document extraction, classification, assistants, and agents. The role blends software engineering and model evaluation: production readiness includes measurable quality, repeatable tests, and clear trade-offs in cost, latency, and reliability. It also involves presenting work to clients and translating ambiguous needs into practical solutions.

Responsibilities - Design, build, deploy, and maintain LLM-powered applications, using prompts, retrieval, fine-tuning, conventional models, or deterministic code as the problem requires. - Create evaluation datasets, grading criteria, baselines, and regression thresholds; use results to guide releases and improve model quality. - Develop production-grade Python services and ship them through tested, reproducible CI/CD workflows. - Build retrieval systems and integrations, measuring retrieval quality and selecting appropriate tools or orchestration frameworks. - Manage deployment and operational trade-offs, including security, privacy, request cost, latency, caching, routing, and streaming. - Lead technical discovery and live demonstrations, explain trade-offs to non-technical stakeholders, and share reusable engineering assets.

Requirements - Five or more years of professional software engineering experience with ownership of production systems, plus at least two years delivering LLM-powered systems. - Advanced Python skills, including testing, asynchronous programming, typing, packaging, and dependency management. - Practical experience designing and running model evaluations, building retrieval-augmented generation systems, and using an LLM orchestration framework. - Hands-on experience with MCP and sound judgment about when to expose capabilities through tools or servers. - Experience deploying AI or ML workloads to a cloud platform, and using Git, automated tests, reproducible builds, and containers. - Understanding of LLM security and privacy concerns, cost and latency controls, and ability to work autonomously in a remote, client-facing team with strong spoken and written English.

Nice to have - LLM evaluation or observability tools, LLM-as-judge methods, fine-tuning, PyTorch, classical ML/NLP, or MLOps experience. - Experience serving open-weight models, building voice or real-time interfaces, or working with legal, financial, or other regulated documents. - Consulting delivery experience, including discovery, demonstrations, or pre-sales; public examples of technical work are also valued.

Benefits and work setup - Fully remote work, available from anywhere, with a monthly service fee in USD. - Twenty business days of paid agreed time off per year, plus six local holidays and caregiver leave.