8 hours ago
San Francisco, CA, USAMid Level
Base Salary
$130k - $220k/yr
Responsibilities
- Design and build response-generation pipelines with routers, orchestrators, and synthesizers.
- Develop knowledge-retrieval systems that ground agent responses in accurate, current data.
- Build reliable multi-step tool orchestration workflows and handle third-party API failures.
- Create evaluation frameworks using real conversations to measure quality and detect regressions.
- Implement observability and tracing for complex agent pipelines.
- Design no-code controls and interfaces for support leads to adjust agent behavior.
- Extend the platform to voice conversations with real-time transcription and tool use.
- Optimize model routing, caching, and prompts for speed and cost.
- Own infrastructure, data pipelines, integrations, and API work supporting the platform.
Requirements
- At least 3 years of experience building and shipping production systems used by real users.
- Experience designing and debugging production AI/ML systems, including retrieval pipelines, prompt engineering, or model orchestration.
- Proficiency across multiple stack layers, including backend, frontend, infrastructure, or data pipelines.
- Experience building and evaluating retrieval systems such as knowledge bases, ranking, context validation, or semantic search.
- Experience with multi-step workflows, tool orchestration, or function-calling systems.
- Experience building observability, tracing, or debugging tools for complex pipelines.
- Proficiency in at least one backend language such as Python, Go, Rust, or Java.
- Experience integrating third-party APIs and handling production failure modes.
- Candidates must be authorized to work in the US without visa sponsorship.
- Experience with real-time or low-latency systems, voice and transcription, model routing, or configuration interfaces for non-technical users is a bonus.
Benefits
- Comprehensive health insurance.
- 401(k) match.
- On-site work in San Francisco, California; not a remote role.
