Software Engineer, Frontier Data Products

Mercor•San Francisco, CA

1d•$160,000 - $500,000•Onsite

About The Position

Mercor is hiring a Software Engineer for its Frontier Data Products team. The team builds production systems that capture, coordinate, and validate expert work and AI-generated work at scale, directly between customer requests and deliverable outputs. The systems are long-running and stateful: jobs can span days, combine automated steps, model inference, and expert review, and allow completed steps to be reopened or changed. The role focuses on durable orchestration and distributed state machines rather than conventional pipelines, with early engineers shaping the architecture of a strategically central product area. The engineer will design and operate backend services for workflows involving automated processing and expert reviewers, build orchestration primitives for retries and recovery, integrate model inference while preserving human oversight and debuggability, and develop APIs and tooling for product, operations, and ML teams. The role owns reliability and observability end-to-end, including for failures that could corrupt results. Work involves translating ambiguous product, operational, and ML constraints into clean system design and debugging complex workflows across automation, models, and human review. Mercor says it works in person five days a week in San Francisco, New York, and London. The posting also contains separate company descriptions referring to a five-day in-person schedule in San Francisco. It describes the company as profitable and gives differing funding and valuation descriptions in its company overview sections.

Requirements

Production backend experience.
Strong system-design instincts, including service boundaries and decisions about where complexity belongs.
Practical fluency with distributed-systems concepts and tools, including asynchronous workflows, queues, idempotency, retries, and long-running jobs.
Ability to turn ambiguous product, operational, and ML constraints into clean, debuggable systems.
Comfort working with Python, AWS, and Postgres.

Nice To Haves

Experience with Temporal or similar workflow engines.

Responsibilities

Design services and state models for multi-stage workflows that distribute work across automated processing and expert reviewers and reconcile results.
Build orchestration primitives, including retries, failure recovery, idempotency, and auditable state transitions, for long-running jobs that may be partially redone.
Integrate model inference into production workflows while maintaining debuggability and human oversight.
Build APIs and tooling that enable product, operations, and ML teams to operate, debug, and trust systems at scale.
Own reliability and observability for workflows, including prevention and detection of silent failures that could corrupt results.
Work with product, operations, and ML teams to translate constraints into system designs.
Own systems end-to-end, including design, shipping, operation, and improvement.
Debug complex stateful workflows spanning automated steps, model calls, and human reviewers.