The opportunity
The Search Product Infrastructure team builds the systems that power search experiences across ChatGPT. We partner with teams developing models, operating inference infrastructure, building specialized search experiences, and maintaining search indexes to bring advances in models and retrieval into production.
What you'll do
Design and evolve the services that coordinate search classification,: retrieval, ranking, and model inference, working closely with researchers and partner engineering teams to bring new capabilities into production.
Improve end-to-end latency, throughput, and infrastructure efficiency through: profiling, caching, request routing, and capacity planning, making informed tradeoffs between search quality, reliability, and compute cost.
Build experimentation tooling and automation to run reproducible A/B tests,: define success metrics, and measure product impact. Use shadow traffic and load testing to validate system behavior, estimate capacity needs, and support safe production rollouts.
Own production reliability through observability, resilient fallback: behavior, incident response, and automation that improves launch safety and reduces operational toil.
Build search APIs and tool interfaces that enable models and agents to: retrieve information reliably while respecting access controls and preserving source attribution.
Have significant experience designing, building, and operating large-scale: distributed systems, with depth in performance, reliability, or resource efficiency.
What they're looking for
- Bring experience in one or more of search, information retrieval, ML: infrastructure, inference serving, or other high-throughput online systems.
- Have strong programming and systems debugging skills, and are comfortable: working across languages and unfamiliar parts of a production stack.
- Can turn ambiguous product or research needs into clear technical plans,: align partners across teams, and carry projects through deployment and measurable results.
- Enjoy learning across systems and ML, investigating unfamiliar problems, and: sharing technical decisions and lessons clearly with others.
