15 hours ago
San Francisco, CA, USAMid Level
Base Salary
$130k - $220k/yr
Responsibilities
- Design response-generation pipelines covering intent routing, orchestration, context assembly, tool selection, and escalation decisions.
- Build retrieval systems that ground responses in accurate and current customer information.
- Develop reliable multi-step tool workflows and handle third-party API failures.
- Create conversation-based evaluation frameworks and observability and tracing systems for agent behavior.
- Build interfaces and controls for non-technical teams to adjust agent behavior.
- Extend the platform to voice with real-time transcription and tool use.
- Improve model speed and cost through routing, caching, and prompt refinement while contributing to infrastructure, data pipelines, integrations, and APIs.
Requirements
- At least 3 years of experience building and shipping production systems used by real people.
- Hands-on experience designing and debugging production AI or machine learning systems, including retrieval pipelines, prompt engineering, or model orchestration.
- Experience building and evaluating knowledge retrieval systems such as semantic search, ranking, or context validation.
- Proficiency in a backend language such as Python, Go, Rust, or Java, with the ability to work across the stack.
- Experience with multi-step workflows or function calling, third-party integrations, production failure handling, observability, and system evaluation.
- Product engineering ownership, ideally in a startup or as a product-focused founder, with a record of shipping production products.
- Experience with real-time or low-latency systems, voice or audio processing, user-facing configuration tools, or model cost and latency optimization is useful.
- Five or more years of experience is a plus.
Benefits
- Health insurance.
- 401(k) match.
- 0.20% to 1.20% equity.
- On-site work in San Francisco, California.