Agentic AI Engineer

Job Title: Staff Engineer — Agentic AILocation: San Francisco, CA (On-site / In-office)About The CompanyWe are a well-funded, high-growth startup based in San Francisco, building advanced AI-driven software designed to automate complex, multi-step workflows used by technical professionals. The company is early-stage, moving quickly, and focused on delivering real production value—not research demos—to enterprise customers.This is a small, senior team where technical leaders have significant ownership, direct access to executives, and accountability for outcomes that directly impact customer adoption and business success.Job SummaryWe are looking for a Staff Engineer — Agentic AI to lead the design, execution, and evolution of our core agent intelligence layer. This role is responsible for building AI agents that take user intent and reliably execute real-world, multi-step workflows under cost, performance, and reliability constraints.You will operate as the technical lead for agentic AI, working closely with engineering, product, user research, and domain experts. Success in this role is measured by agent task completion, cost efficiency, and real-world workflow coverage, not by architectural novelty.Key ResponsibilitiesAgent Performance & EvaluationOwn agent task success rateDefine the primary success metrics for agent performance and systematically improve completion rates on real user workflows.Set and enforce cost and token budgetsEstablish per-task cost constraints, track efficiency metrics, and ensure agent behavior is commercially viable at scale.Build robust evaluation infrastructureDesign reproducible, adversarial evaluation frameworks grounded in real user stories rather than synthetic benchmarks, with strong regression detection.User Story Mapping & Workflow CoverageLead workflow discovery and validationPartner with user researchers and domain experts to document how users actually work, validating assumptions through interviews and field feedback.Expand end-to-end workflow coveragePrioritize and deliver increased agent coverage across the highest-value workflows, balancing technical feasibility with customer impact.Translate workflows into measurable testsEnsure every validated user story becomes a benchmarked test case, closing the loop between research and performance.Technical LeadershipOwn the agent architectureMake and own foundational decisions around tool usage, state management, orchestration, error recovery, model routing, and context handling across multi-step workflows.Lead and mentor AI engineersAct as a player-coach—writing code, reviewing designs, raising engineering standards, and unblocking execution for a small senior team.Collaborate cross-functionallyWork closely with platform and integration teams to define platform capabilities, with product teams on user interaction design, and with customers during pilots to incorporate real-world feedback.RequirementsEducationBachelor’s degree in Computer Science, Engineering, or equivalent practical experience.Experience7+ years of professional software engineering experience.2+ years building agentic AI systems that take actions in real environments (tool use, stateful workflows, error handling, cost constraints).Technical SkillsDeep experience with LLM application architecture beyond prompt design, including model selection, orchestration, retrieval strategies, context management, and tool-calling frameworks.Strong proficiency in Python and modern AI tooling ecosystems.Experience designing and operating evaluation and benchmarking systems for AI agents (task completion, efficiency, regression tracking).Leadership & ExecutionProven technical leadership of small teams, setting direction and making architecture decisions (formal people management not required).Strong bias toward measurement, iteration, and shipping reliable systems over producing elegant demos.Preferred QualificationsExperience with agents operating in constrained or production environments (cost limits, latency requirements, reliability expectations).Background in desktop automation, systems integration, or non-web-based tooling.Experience contributing to internal or external benchmarks or evaluation frameworks for AI systems.Familiarity with enterprise deployment or security constraints.Open-source or published work related to agentic AI systems.BenefitsCompetitive Salary: Staff-level compensation aligned with ownership and impact.Health and Wellness: Comprehensive medical, dental, and vision coverage.Work Environment: In-office role in San Francisco with close collaboration across engineering and product.Professional Development: Opportunity to define foundational AI systems at a fast-growing startup.Additional Perks: High ownership, direct executive visibility, and influence over long-term technical direction.