Staff Software Engineer, AI/ML Infrastructure

Thumbtack helps millions of people confidently care for their homes through personalized guidance, AI tools, and a marketplace for local service businesses. The Staff Software Engineer will provide technical leadership while designing, building, and scaling AI/ML infrastructure and evaluation platforms, including LLM gateways, model training and serving, data workflows, orchestration, observability, and regression evaluation systems. The role also partners across engineering and applied science teams to define roadmaps and deliver high-impact production systems.

Responsibilities:

• Build and evolve core AI platform capabilities that enable teams to develop, run, and scale GenAI-powered applications across Thumbtack
• Contribute to the design, development, and deployment of scalable tools and infrastructure to support the efforts of our applied scientists, including traditional ML model training and serving systems, feature and data workflows, CI/CD, orchestration, deployment, and evaluation tooling
• Work hands on across the stack, from backend services and execution infrastructure to integrations with AI models and tooling
• Partner with senior engineers to evaluate next-generation AI infrastructure frameworks and tools that help product teams harness advances in AI
• Drive projects to completion with a strong focus on business impact and measurable outcomes
• Solve complex technical problems and stay up to date with advances in this rapidly evolving space

Requirements:

• 10+ years of professional software engineering experience, with a track record of technical leadership at staff level or equivalent scope, setting direction for systems spanning multiple teams, driving alignment across teams, and seeing them through to production
• 3+ years building AI/ML infrastructure in production including model serving, feature platforms, training or orchestration systems, or ML observability
• Hands-on experience with LLM-powered systems in production, including at least some of the following: LLM gateways or proxies, prompt and model lifecycle management, RAG or agentic architectures, and cost optimization for LLM workloads
• Experience building or operating AI evaluation systems: offline eval pipelines, LLM-as-judge approaches and their calibration against human judgment, regression testing for model or prompt changes, or A/B measurement of AI features
• Deep experience designing, building, and operating reliable distributed systems, including ownership of production services at meaningful scale, on-call responsibility, and incident leadership
• Demonstrated ability to use AI coding tools in day-to-day workflows and to validate, critique, and refine AI-generated output