Job description

ABOUT THE ROLE A well-funded, early-stage B2B SaaS startup in the AI-powered sales automation space is hiring its first Founding Engineer. You'll join the CTO as a core member of a tiny, high-caliber team and help set the technical culture for everyone who comes after. The company builds self-improving conversational agents that run product demos 24/7 — adapting to each buyer, handling technical questions, and getting smarter with every conversation. Growth has tripled month-over-month and demand is accelerating. WHAT YOU'LL WORK ON There are no strict lanes here. You'll own problems across the full stack, from LLM orchestration to embeddable frontend widgets to eval pipelines. Key technical pillars include: - The Agent: LLM orchestration, conversation flow, tool use, and voice. Agents must understand unfamiliar products and explain them clearly to prospects with varying levels of context. - The Experience: Embeddable, generative UI that adapts in real-time — fast-loading, interactive, and responsive across products and screen sizes. - The Platform: Customer-facing tooling for configuring agents, managing flows, reviewing conversations, and measuring performance. - Evals & Self-Improvement: Pipelines that measure agent quality and feed learnings back automatically, so demos get smarter without manual intervention. - Infrastructure: Zero cold starts, instant agent response, and versioning systems that let customers preview changes before they go live. DAY-TO-DAY EXAMPLES - Debug conversation logs, trace where an LLM lost the thread during a pricing objection, and ship a fix to the orchestration layer by lunch. - Prototype a generative UI widget in an afternoon based on a whiteboard sketch and have something demo-able by end of day. - Trace a quality regression to a prompt change, roll it back, and add an automated test to prevent recurrence. - Design agent versioning infrastructure from scratch so customers can safely preview updates. WHAT WE'RE LOOKING FOR Required: - Proven experience designing and building LLM-based conversational agents — orchestration, tool use, prompt engineering, and conversation flow. - Production backend development experience in Python (services, APIs, backend infrastructure). - Frontend experience with React — building interactive, adaptive UIs and rapid prototyping. - Demonstrated ability to deliver full-stack features end-to-end: frontend, backend, integration, and deployment. - Experience building evaluation pipelines, automated regression tests, observability tooling, and CI/CD workflows. - Experience designing deployment and versioning infrastructure for agents (feature flags, preview environments, versioned rollouts). - Experience with voice-enabled or browser-based agents (STT/TTS, WebRTC, or browser automation) or equivalent voice/browser + LLM integration experience. - Experience deploying and operating production services on cloud platforms (AWS or GCP) with Docker; Kubernetes or similar orchestration a plus. - Strong product sense: ability to rapidly prototype, iterate on feedback, and prioritize high-impact work in a fast-moving environment. - Experience at an early-stage startup or demonstrated ability to thrive in ambiguous, high-velocity environments. - Strong written and verbal communication skills — comfortable debugging issues with customers and documenting design decisions clearly. - Willingness to be available for urgent production issues and participate in on-call or incident response rotations. Nice to Have: - Familiarity with reinforcement learning workflows (RL, RLHF, reward modeling) or experimentation frameworks applied to agent improvement. COMPENSATION & BENEFITS - Salary: $180,000 – $230,000 USD annually - Early-stage equity commensurate with founding engineer role - Visa sponsorship is not available — candidates must be legally authorized to work in the United States LOCATION This role is fully on-site in San Francisco, CA, five days per week. Remote or hybrid arrangements are not available for this position.

Requirements

Must-haves

There are no strict lanes here. You'll own problems across the full stack, from LLM orchestration to embeddable frontend widgets to eval pipelines. Key technical pillars include:
The Agent: LLM orchestration, conversation flow, tool use, and voice. Agents must understand unfamiliar products and explain them clearly to prospects with varying levels of context.
The Experience: Embeddable, generative UI that adapts in real-time — fast-loading, interactive, and responsive across products and screen sizes.
The Platform: Customer-facing tooling for configuring agents, managing flows, reviewing conversations, and measuring performance.
Evals & Self-Improvement: Pipelines that measure agent quality and feed learnings back automatically, so demos get smarter without manual intervention.
Infrastructure: Zero cold starts, instant agent response, and versioning systems that let customers preview changes before they go live.
Debug conversation logs, trace where an LLM lost the thread during a pricing objection, and ship a fix to the orchestration layer by lunch.
Prototype a generative UI widget in an afternoon based on a whiteboard sketch and have something demo-able by end of day.
Trace a quality regression to a prompt change, roll it back, and add an automated test to prevent recurrence.
Design agent versioning infrastructure from scratch so customers can safely preview updates.
Proven experience designing and building LLM-based conversational agents — orchestration, tool use, prompt engineering, and conversation flow.
Production backend development experience in Python (services, APIs, backend infrastructure).
Frontend experience with React — building interactive, adaptive UIs and rapid prototyping.
Demonstrated ability to deliver full-stack features end-to-end: frontend, backend, integration, and deployment.
Experience building evaluation pipelines, automated regression tests, observability tooling, and CI/CD workflows.
Experience designing deployment and versioning infrastructure for agents (feature flags, preview environments, versioned rollouts).
Experience with voice-enabled or browser-based agents (STT/TTS, WebRTC, or browser automation) or equivalent voice/browser + LLM integration experience.
Strong product sense: ability to rapidly prototype, iterate on feedback, and prioritize high-impact work in a fast-moving environment.
Experience at an early-stage startup or demonstrated ability to thrive in ambiguous, high-velocity environments.
Strong written and verbal communication skills — comfortable debugging issues with customers and documenting design decisions clearly.
Willingness to be available for urgent production issues and participate in on-call or incident response rotations.
This role is fully on-site in San Francisco, CA, five days per week. Remote or hybrid arrangements are not available for this position.

Nice-to-haves

Experience deploying and operating production services on cloud platforms (AWS or GCP) with Docker; Kubernetes or similar orchestration a plus.
Familiarity with reinforcement learning workflows (RL, RLHF, reward modeling) or experimentation frameworks applied to agent improvement.

Application prep

Use the role’s language and application details while tailoring your materials.

Keywords to mirror

aillmgosketchpythonreactci/cdawsgcpdockerkubernetes

Platform: ashby
Effort: quick