Associate AI Engineer, Figure

• Strong programming fundamentals and practical Python experience
• Working knowledge of SQL and structured data analysis
• Ability to break ambiguous problems into testable components
• Comfort reading unfamiliar code and tracing system behavior
• Strong written communication and attention to detail
• Interest in AI behavior, analytics, backend systems, and product workflows
• A habit of validating assumptions and measuring outcomes
• Ability to collaborate with engineers and nontechnical domain experts
• Desire to grow into a full-stack production AI engineer
• Experience building an API, backend service, automation, data pipeline, or internal tool
• Experience evaluating LLMs, conversational agents, or other probabilistic systems
• Knowledge of experimental design, statistics, causal inference, or applied econometrics
• Experience testing nondeterministic systems
• Familiarity with retrieval-augmented generation, agent tools, workflow engines, or model evaluation
• Ability to distinguish problems requiring deterministic logic from those suited to an LLM
• Experience translating domain or policy requirements into software behavior
• A PhD is welcome but not required. Candidates with unconventional backgrounds are encouraged to apply if they demonstrate strong analytical reasoning and practical software-building ability

Desirable

• We're hiring an Associate AI Engineer to help operate and improve our conversational AI systems while developing into a full-stack AI engineer
• Initially, approximately half of your work will focus on the technical operation of AI-powered chat, voice, and customer-support services
• You will bring engineering rigor to agent workflows, API integrations, testing, analytics, knowledge retrieval, security, and production releases
• The other half will consist of hands-on projects across our broader AI stack. With coaching from senior engineers, you will build Python services, APIs, data pipelines, evaluations, and internal tools supporting production AI systems
• This role is well suited to someone who combines analytical depth with practical programming ability. You may come from software engineering, economics, another quantitative social science, data science, or a similarly rigorous field. We care more about demonstrated reasoning and engineering ability than a particular degree or conventional career path
• This is not a content-management or prompt-writing position. You will help operate conversational AI as production software: tested, measurable, secure, observable, and carefully governed
• Improve AI-powered chat and voice services supporting customer-service and lending workflows
• Build and maintain agent workflows, API-integrated tools, guardrails, retrieval rules, and escalation paths
• Review workflows for logic errors, unsafe behavior, routing conflicts, and unnecessary model context
• Establish disciplined development, staging, testing, review, and production-release practices
• Partner with domain experts who own customer experience, policies, and lending operations
• Translate business knowledge into reliable system behavior without becoming the sole owner of that knowledge
• Determine whether problems are best solved through workflow logic, deterministic code, retrieval, an LLM, or a product change
• Build automated tests and simulations for conversational workflows
• Define and measure containment, escalation, answer quality, task completion, and customer-impact metrics
• Develop tagging, monitoring, and quality-assurance systems for production conversations
• Analyze failures and turn production evidence into prioritized improvements
• Build pipelines that make conversational data available to analytics and reporting systems
• Design controlled experiments and incremental rollouts that measure business outcomes
• Build Python services and APIs that expose AI capabilities
• Develop integrations between AI systems and Figure's internal services
• Build data and evaluation pipelines using Python and SQL
• Contribute to internal tools and user interfaces
• Learn Figure's deployment, monitoring, containerization, and CI/CD practices
• Gradually own increasingly substantial production engineering projects
• Inventory and review conversational tools that access internal or third-party APIs
• Ensure sensitive operations use deterministic validation instead of unreliable model judgment
• Support appropriate handling of customer data, credentials, PII, and regulated workflows
• Maintain auditability for workflow changes and production behavior
• Escalate security, compliance, and reliability concerns using sound technical reasoning
• What Success Looks Like:
• Within your first six months, you will:
• Understand Figure's conversational AI workflows and integrations
• Strengthen testing, monitoring, security, and release discipline
• Resolve concrete workflow reliability issues
• Produce defensible reporting on conversational-system performance
• Ship at least one production integration, service, or internal tool
• Demonstrate increasing independence across Python, APIs, data, evaluation, and deployment
• Over time, the role will shift toward broader full-stack AI engineering as the conversational systems become better governed and operational responsibilities become more distributed