
Posted 5 days ago
Technical Program Manager
AI Summary
Leads data and evaluation programs for AI research teams, from requirements through production, QA and delivery, while managing vendors and improving data quality and pipelines.
About this role
About the Role
Own complex data and evaluation programs for AI research teams and internal partners. This role brings together technical program management, data operations, and vendor coordination to help ensure evaluation data is high quality and useful for improving AI systems.
What You'll Do
Lead data and evaluation programs from scoping and requirements gathering through production, quality assurance, delivery, and retrospectives.
Turn ambiguous requests into clear specifications with milestones, owners, dependencies, and acceptance criteria.
Assess and improve data quality using quantitative metrics and qualitative review, including task coverage, difficulty, and reward distributions.
Manage external vendors throughout the delivery lifecycle and improve pipelines as programs scale.
Improve workflows, documentation, automation, and cross-functional handoffs for data production.
Partner with research, platform, and marketplace teams to turn feedback into product improvements.
What We're Looking For
At least 3 years of experience owning complex technical programs in AI/ML, data operations, support engineering, applied research, forward-deployed engineering, or a related cross-functional field.
Startup experience as a Technical Program Manager or Strategic Program Lead, with the ability to work independently in a fast-paced environment.
Experience translating ambiguous technical requirements into executable plans and maintaining or improving complex data quality processes.
Strong judgment in evaluating datasets, moving between aggregate metrics and individual examples.
Experience managing vendors, scaling delivery pipelines, and supporting AI research teams or technical customers with evolving requirements.
Familiarity with RLHF, reinforcement-learning environments, agent evaluations, human-in-the-loop data pipelines, annotation, or expert data collection.
Clear written and verbal communication, sound prioritization, and comfort navigating changing requirements.
Compensation & Benefits
Base salary is $130,000 to $225,000 USD annually. Visa sponsorship and relocation support are available for qualified full-time candidates.
Location
On-site in San Francisco, California. Remote candidates may be considered if they can maintain 70% to 80% time zone overlap with San Francisco or Singapore.
Skills
Explore related jobs
More jobs at Clera
Similar Agent Evaluations jobs
- A
AI Agent EvaluatorAnyone AI · Argentina - Fully Remote - N
Agent Evaluation and Self-Improvement EngineerNVIDIA USA · US, CA - G
Software Engineering Director, Agentic EvaluationsG2 · Remote (US) - Z
Director, Enterprise Quality Engineering and Agentic EvaluationZayo Group LLC · United States - B
Applied Scientist, Agent Evaluation & Adaptive Model RoutingBitdeer Technologies Group · Singapore, Singapore - ASenior Technical Product Manager – Agent Evaluation & ObservabilityAmerican Express · New York, NY
Jobs in San Francisco
- P
Account Manager - Banking and WealthPlaid · San Francisco HQ - P
Internal Audit AnalystPlanet · San Francisco, California - P
Expert Support Manager (Contract, San Francisco)Patronus AI, Inc. · San Francisco - J
Founding Business Development RepresentativeJudgment Labs · San Francisco - F
Offensive Security EngineerFinix · San Francisco - S
ASC Nurse ManagerSpring Fertility · San Francisco, California
Browse these categories
Market data for this role
All reports →- SeriesRole reportsOne role family at a time: how many openings, what changed this week, who is hiring, what it pays.
- SeriesSalary reportsWhat employers publish in job postings, by level and workplace. Not self-reported pay.
- Market overviewState of tech hiring, September 2026: up 4.8%Tech hiring rose 4.8% month over month in September 2026, with 411,122 new listings. Customer support and account executive roles led the growth.