Product Operations Manager, Research Services

Mercor•San Francisco, CA

10h•$140,000 - $210,000•Onsite

About The Position

As a Product Operations Manager, Research Services, you'll drive operational excellence to support Mercor's full-service research partnerships. You will prioritize and manage delivery across various strategic customer segments to support model evaluations, diagnose model capability gaps, and fulfill post-training demand to better support our most strategic customers. This is a highly technical, internally-facing role in the Product Operations job family at the intersection of Product, Engineering, Data Production, GTM, Finance, and Strategic Partnerships.

Requirements

3+ years in Product Operations, Product Management, Technical Program Management, Research Operations, Solutions Engineering, or a similar technical, customer-facing role.
Experience operating in an ML research or AI infrastructure environment strongly preferred.
Fluent in how modern LLM evaluation works, including harness design, verification, agentic rollouts, and contamination and fairness pitfalls.
Working understanding of post-training (SFT, preference ranking, RL) and of inference economics.
Familiar with coding agents; comfortable in SQL and Python without agent assistance.
Can plan capacity against an uncertain demand curve, hold partners to an SLA, and weigh cost against reliability without needing to escalate every call.
Thrive in ambiguous environments and take ownership from problem definition through execution. Comfortable being the person accountable for a date.
Can hold your own in a technical conversation with AI researchers and write a status update an executive can act on. Excellent written and verbal communication.
Strong project management and cross-functional coordination skills. Passion for fixing problems at the source and building repeatable systems rather than one-off solutions.

Responsibilities

Prioritize the Benchmark Portfolio: Own the intake and prioritization of new benchmarks and evaluations supported in our platform. Weigh customer demand, observed capability gaps, differentiation versus public benchmarks, and build cost to inform our priorities.
Support Dataset Demand Generation: Run and share leaderboard evals to drive Mercor’s research brand equity during new model releases.
Manage Training and Inference Partners & Budgets: Own the relationships with our inference providers, compute partners, and external training vendors. Forecast capacity against the delivery calendar, negotiate and track SLAs and rate cards, monitor cost per eval run and per rollout, and qualify new partners before we need them.
Track Customer Delivery: Maintain the operating picture across every active evals, loss analysis, and post-training engagements, highlighting scope, milestones, dependencies, and SLA status. Run the delivery review, escalate slipping commitments before the customer notices, and give research counterparts a straight answer on where things stand.
Shape the Product: Translate recurring delivery friction into product requirements for the eval platform and partner with Engineering to automate the workflows you'd otherwise be running by hand.