Talent Analyst

Luma AI•Palo Alto, CA

1d•Hybrid

About The Position

Talent Analytics & Operations (TAO) exists to make Luma's growth engine intelligent. We build the data and systems that let Luma hire world-class researchers and engineers faster than anyone else in AI. We're a small team of builders, not administrators. You'll be the first dedicated analytics hire, working alongside ops and automation specialists and partnering with recruiting leads across Research, Applied Research, Product Engineering, GTM, and Business Operations. Is this role for you? Yes, if you: Get excited when you find a pattern in messy data that nobody else noticed Have built a dashboard or report that actually changed how someone made a decision Want to define new metrics, not just report on existing ones Prefer shipping something imperfect over planning something perfect Want to own problems end-to-end, not wait for someone to hand you requirements Are energized by ambiguity and fast pace Probably not, if you: Want a well-defined role with clear boundaries and established processes Prefer deep specialization over breadth Need a large team and extensive onboarding to be effective Are looking for a pure data science or ML role Want to work on one type of problem for a long time You start the morning in a working session with the Head of Recruiting, sketching out how to measure "quality of hire." It's a metric everyone wants but nobody agrees on. You're mapping out what signals we could actually capture (hiring manager satisfaction at 30/90 days, time to first meaningful contribution, performance review correlation) and what data infrastructure we'd need to make it real. Mid-morning, you're building a draft schema for a lightweight data warehouse. Our ATS and HRIS don't talk to each other natively, and you're tired of stitching together exports manually. You're evaluating whether we need a proper ETL pipeline or if a simpler solution gets us 80% of the value. After lunch, you're presenting to leadership. Not a status update, but a recommendation: our Research hiring funnel converts 3x better than Product Engineering, and you've got a hypothesis why. You're proposing an experiment to test it. Late afternoon, you're researching how other companies measure recruiter effectiveness. Most of the industry uses activity metrics (screens per week, submits per role). You think that's backwards. You're drafting a framework for outcome-based measurement that could become how Luma thinks about recruiting performance, and maybe something we share externally. You end the day automating a report that used to take someone two hours every Monday. Small win, but it adds up. You're a data person who's figured out how to make insights land. You write SQL without thinking about it. You've built dashboards or reports that people actually used to make decisions. You can explain a data quality issue to a recruiter and a funnel trend to a CEO. You've built things, not just analyzed things. Maybe you've stood up a lightweight data warehouse. Maybe you've designed a schema that made reporting actually work. You understand that good analytics requires good infrastructure, and you're not afraid to build it. You think about measurement as a craft. You're not satisfied with vanity metrics or industry defaults. You want to figure out what actually matters, how to capture it, and how to make it useful. You've probably been frustrated by how most companies measure things. You've worked somewhere fast. Startups, high-growth companies, or scrappy teams where priorities shift and you figure it out. You're not waiting for perfect requirements. You're shipping and iterating. You're comfortable with ambiguity. We're handing you problems, not specs. You'll need to figure out what question we're actually trying to answer, find (or build) the data, and present something useful. You communicate clearly. You can turn a messy analysis into a clean story. You know when to caveat and when to be direct. You're not precious about your work. You'd rather be right than look smart.

Requirements

2-4 years working with data in a business context (analytics, ops, BI, or similar)
Strong SQL skills (complex queries, joins, window functions without Googling)
Experience building dashboards or reports that informed real decisions
Some experience with data infrastructure (databases, pipelines, ETL concepts)
Comfort with messy, imperfect data and the judgment to know what to fix vs. ignore
Experience at a startup or high-growth company (you know the pace)
Clear written and verbal communication

Nice To Haves

Python for data work (pandas, notebooks, API integrations)
Experience designing or maintaining a data warehouse (even a lightweight one)
Experience with recruiting, HR, or people data specifically
Exposure to, ideally worked strongly with, AI-assisted tools and workflows (using Claude, ChatGPT, or similar for data work)
Familiarity with ATS/CRM systems (Gem, Greenhouse, Lever, Ashby, etc.)
Experience with automation tools (Zapier, n8n, Make)
People management experience

Responsibilities

Recruiting analytics and reporting
Build and maintain dashboards that leadership actually uses (weekly, without anyone asking)
Reduce recruiting leadership's manual reporting time by 80%
Enable hiring managers to self-serve pipeline data in real-time
Create the metrics and visualizations the CEO looks at weekly
Data infrastructure
Design and build lightweight data infrastructure (warehouse, pipelines, integrations) to support analytics at scale
Connect fragmented systems (ATS, HRIS, scheduling tools) into a unified data layer
Own data quality, reliability, and documentation so the team can trust what they're seeing
New metrics and frameworks
Define metrics that don't exist yet at Luma (quality of hire, source effectiveness, time to productivity)
Partner with recruiting leads and hiring managers to understand what they need to measure and why
Research industry best practices, then build something better
Create measurement frameworks that could become how Luma thinks about talent, and potentially how the industry does
Insights and recommendations
Translate data into "so what" and "now what" for non-technical stakeholders
Identify trends, anomalies, and opportunities that inform recruiting strategy
Answer the questions leadership doesn't know to ask yet
Automation and workflows
Automate repetitive data pulls, report generation, and cross-system workflows
Build lightweight internal tools when off-the-shelf solutions don't fit
Connect data sources to reduce manual entry and human error