Machine Learning Engineer, Checkr

• We need someone who writes production software, builds with LLMs and APIs as first-class tools, and can tell the difference between working code and AI slop. If you’ve spent the last few years building AI-native software and you care deeply about engineering craft, we want to talk
• We are looking for someone based in San Francisco who has built ML systems in fast-moving, impact-first environments. Less process, more shipping. Less paperwork, more results
• A Bachelor’s or Master’s degree in Computer Science, Mathematics, or a related technical field, or equivalent depth from experience
• 4+ years building software professionally, with at least 2 of those building ML systems that run in production
• Strong Python fluency; you write clean, testable, well-structured code with solid OOP instincts. Not scripts; software
• Hands-on experience using LLM APIs in production systems: prompt engineering, structured outputs, function calling, cost management, and evaluation
• You’ve built and maintained APIs, worked with CI/CD pipelines, and shipped code that other engineers depend on
• Comfortable with distributed systems concepts: queues, async processing, caching, horizontal scaling
• Experience with NLP tasks in production: classification, extraction, entity resolution, summarization
• Comfort with and enthusiasm for AI-assisted workflows; experience using LLMs, code-generation tools, or agentic systems in production or operational contexts is a strong signal
• You can evaluate tradeoffs: fine-tune vs. prompt, hosted vs. self-deployed, classical ML vs. LLM, rule vs. model
• Strong communication skills; you explain technical decisions clearly to engineers and non-engineers alike, without hiding behind jargon
• You use AI tools (Copilot, Claude, etc.) to move faster, but you understand every line they produce. You can spot AI slop and you don’t ship it
• An A-player mindset with a strong bias for action: you raise the bar, move with urgency, stay resilient through ambiguity, and take ownership to deliver meaningful outcomes

Desirable

• We’re hiring an ML Engineer (P2) to build and ship the AI systems that power Checkr’s core products. This role sits on the ML team inside Checkr’s Data & ML organization within Engineering
• Checkr runs millions of background checks a year. The ML team builds the systems that make those checks faster, more accurate, and cheaper to operate: document processing, charge classification, entity resolution, and in-product intelligence
• These are production services that Product Engineering depends on daily
• This is not a research role or a notebook role. You’ll own ML services end-to-end: design them, code them, deploy them, monitor them
• This role sits in the central Data & ML team within the Engineering organization. You will partner daily with Product Engineering, Product, and cross-functional teams. You’ll also contribute to Checkr’s broader AI strategy, including our initiative to deploy our agentic fleet and build scalable context with our semantic layer
• Build and deploy ML/AI services. Design, develop, and ship ML models and AI systems that Product Engineering teams rely on. You write the model code, the API layer, the monitoring, and the tests. Not notebooks; production services
• Design with LLMs and APIs. Use LLM APIs (OpenAI, Anthropic, etc.) as building blocks in production systems. You know when to call an LLM, when to fine-tune, when to use a classical model, and when to write a rule. You think about cost, latency, and quality together
• Ship production software. Write clean, well-structured code with solid OOP, proper abstractions, error handling, and tests. Your code gets reviewed by SWEs and passes. CI/CD is how you work, not something you bolt on at the end
• Partner with product and engineering. Translate business problems into ML solutions. Define API contracts with product engineers. Explain your approach clearly to non-ML partners and leave the room with alignment, not confusion
• Evaluate and iterate fast. Build evaluation frameworks, run experiments, and make data-driven decisions about model and system performance. Ship and iterate; don’t wait for perfect
• Ship AI-powered workflows. Put AI to work on your own processes: automate pipelines, build agentic workflows, and contribute reusable skills and context to Checkr’s agentic platform. The expectation is that our teams operate AI-first