AI-Product Engineering Leader

EY

San Francisco (CA)

Hybrid

USD 157,000 - 262,000

Full time

4 hours ago

Be an early applicant

Get a reply from this employer

Your resume and cover letter, tailored to exactly what they’re hiring for.

Get past ATS filters

Job summary

EY is pursuing Challenger, a program to shape the firm’s next generation of AI-powered services. This role focuses on turning emerging AI capabilities into durable products, working end-to-end from user conversations to production services, with a strong emphasis on speed, accessibility, and maintainability.

The candidate will join Challenger Core and collaborate with small, dedicated teams to build reusable tools and guardrails that accelerate delivery.

Qualifications

7+ years of relevant and applicable experience.
End-to-end delivery: take customer- or employee-facing software from concept through launch, learning, and adoption.
Technical breadth: full-stack capability across modern frontend, backend, API, data, and cloud technologies.
Applied AI experience: hands-on experience building products with large language models, retrieval, tools, or comparable AI application patterns.
Evaluation experience: evidence of automated testing and behavioral evaluation for systems with variable or probabilistic outputs.
Live-service ownership: telemetry, debugging, reliability, incident response, recovery, and performance improvement.
User-centered practice: experience working directly with users, designers, and product leaders to turn evidence into better product decisions.
Engineering craft: a consistent record of clear specifications, maintainable code, thoughtful review, and sound technical trade-offs.
This role requires up to 50% travel and flexibility to go into the office in a hybrid capacity.

Responsibilities

Start with the user. Work directly with users and domain experts to understand real workflows, observe friction, and identify where AI can create distinctive value.
Test assumptions quickly. Turn product hypotheses into concise specifications, prototypes, and small releases that generate useful evidence early.
Own complete product slices. Build coherent verticals across the experience, services, data, AI behavior, and enterprise integrations rather than optimizing a single layer.
Shape AI behavior. Design instructions, context, retrieval, structured outputs, and tool use, choosing adaptation or tuning only when evidence supports it.
Build the evidence system. Create representative tasks and automated evaluations that show whether the product is useful, correct, safe, accessible, and improving.
Connect outcomes to behavior. Instrument user journeys so the team can relate adoption and task success to system behavior, latency, failures, and cost.
Engineer for responsible use. Build clear permission boundaries, secure data handling, human checkpoints, graceful degradation, recovery, and rollback into the experience.
Own production quality. Release incrementally, monitor live behavior, respond to incidents, and turn failures into durable product and engineering improvements.
Make practical trade-offs. Balance experience, quality, speed, reliability, and cost, using deterministic software when it is the more dependable choice.
Contribute reusable patterns. Share proven components and evaluation assets with the broader platform while remaining accountable for their performance in the product workflow.