AI Engineer

ServiceTitan is hiring an AI Engineer (onsite in California) to build safe, observable, production-grade AI agent infrastructure for the Agent OS platform.

Responsibilities

• Build the runtime for role-specific agents, including tool use, delegation, pause/resume, durable checkpoints, retries, and failure recovery
• Implement deterministic controls around non-deterministic reasoning, including governed reads, proposed writes, precondition checks, business invariants, scoped permissions, idempotency, audit trails, and rollback
• Create tenant-scoped context assembly with retrieval, freshness controls, provenance, transcripts, artifacts, tool results, and replayable evidence
• Develop human-in-the-loop gates, approval thresholds, reversibility, tenant policy enforcement, and audit history for financial, contractual, dispatch, warranty, and compliance-sensitive workflows
• Build offline and online evaluation systems, including scenario libraries, simulation, trajectory review, regression detection, and cost and latency telemetry with autonomy promotion gates
• Help product teams package prompts, tools, context requirements, policies, evaluations, rollout controls, ownership, and rollback into governed capabilities for owners, CSRs, dispatchers, technicians, managers, and back-office teams
• Make platform-level tradeoffs across latency, cost, quality, structured outputs, caching, fallback behavior, provider choice, and model routing
• Design and implement core Agent OS platform services and agent execution workflows
• Write production code and review implementation details from other engineers
• Inspect traces, debug failures, and improve production behavior for agent reliability
• Design evaluation scenarios and regression suites for agent workflows
• Work through real agent failure modes such as stale context, wrong tool calls, missing permissions, unsafe actions, poor retrieval, latency spikes, and cost regressions
• Partner with domain teams to convert agent use cases into reusable platform patterns
• Help define platform contracts for tools, actions, approvals, context, memory, evidence, and evaluation
• Provide technical direction while staying grounded in what can be shipped quickly and safely
• Communicate clearly with engineers, product managers, architects, security partners, and engineering leadership

Requirements

• Experience building production code and designing reliable APIs, workflows, tools, and services for agent execution
• Ability to design deterministic controls for non-deterministic reasoning, including governance, audit trails, rollback, and idempotency
• Experience with context and evidence systems, including tenant scoping, provenance, transcripts, artifacts, tool results, and replayable evidence
• Experience designing evaluations and regression suites, including offline and online evals, simulation, trajectory review, and telemetry for cost and latency
• Ability to debug agent failures using traces, improve production behavior, and handle failure modes such as stale context, wrong tool calls, missing permissions, unsafe actions, poor retrieval, latency spikes, and cost regressions
• Collaboration experience with product, architecture, security, and engineering leadership to ship governed platform capabilities

Compensation

• USD 137,900 - 221,400 per year

Benefits

• Flexible time off with ample learning and development opportunities
• Comprehensive onboarding program
• Leadership training for Titans at all levels
• Bonusly, peer-nominated awards, and more
• Company-paid medical, dental, and vision (with 100% employer paid options and 90% coverage for dependents)
• FSA and HSA
• 401k match
• Telehealth options including memberships to One Medical
• Parental leave and support
• Up to $20k in fertility services (IUI and IVF), surrogacy, and adoption reimbursement
• On demand maternity support through Maven Maternity
• Free breast milk shipping through Maven Milk
• Pet insurance
• Legal advisory services
• Financial planning tools

Human-first approach: Being human isn’t about checking every box; if you have the skills but hesitated based on background, apply anyway.

Use of AI technology: Automated and AI-assisted tools may support parts of the recruitment process, but AI tools are not used to make hiring decisions; hiring decisions are made by the hiring teams.

What we offer: Flextime, recognition, and support for autonomous work; holistic health and wellness benefits; support for Titans at all stages of life.