Senior Software Engineer, Agents

Ellipsis Health; San Francisco (CA); On-site; USD 160,000 - 210,000; Full time

Benefits offered by this job

Health, dental, vision insurance

401(k) plan

Learning budget / conferences

Job summary

Ellipsis Health in San Francisco is hiring a Senior Software Engineer to build AI agents that patients actually talk to. You will own agents end to end—from conversation design and prompts to tool calls and production APIs—shipping multiple agents in production within the first months.

You will collaborate with SVP of Engineering and cross-functional teams to deliver robust, safe, fast, and scalable solutions for healthcare workflows, including scheduling, refills, and billing.

Qualifications

• 5+ years of professional software engineering experience, including 1+ year shipping LLM-powered apps.
• Strong Python and/or TypeScript, backend APIs and data, with frontend basics to ship a full feature.
• Hands-on prompting, tool use and function calling, RAG, structured outputs, and agent orchestration.
• Experience designing evals and using data to decide what to fix next.
• Product sense: patient outcomes matter, not just model metrics.
• Clear written and verbal communication with technical and non-technical teammates.

Responsibilities

• Design, build, and ship AI agents end to end, including conversation flows, prompts, tool and function calling, retrieval, memory, and the backend services agents depend on.
• Build evaluation systems (offline test suites, simulated conversations, human review loops, and production monitoring) to ensure an agent is good before and after it ships.
• Improve instruction following, tool-calling reliability, latency, and cost across LLM and speech stack.
• Help design and build next-gen voice agent platform, choosing models and architecture for quality, latency, and cost.
• Partner with clinical, product, and customer teams to turn workflows (scheduling, refills, billing, post-discharge check-ins) into reliable agent behavior.
• Dig into real conversations and failure cases, and close the loop with prompt, tool, model, or product changes.
• Build internal tools to observe, monitor, and analyze agent behavior at scale to identify failure modes.
• Track what's new in models and tooling, and bring useful parts into production quickly.

Skills

Python/TypeScript

LLM apps

Prompting

Tool calling

Eval design

Communication

Tools

Temporal

Job description

Ellipsis Health builds AI products that take on healthcare's staffing shortages and administrative burden, pairing conversational AI with our patented voice biomarker technology to help deliver better healthcare for everyone. Our voice agents talk with patients every day: booking and rescheduling appointments, answering billing questions, handling medication refills, and checking in after a hospital visit.

This role sits on our provider engineering team, building the agents that health systems, provider groups, and FQHCs deploy with their patients. You'll work across the provider teams within Ellipsis, partnering with product, clinical, deployment, and commercial teams to take new agent workflows from idea to production.

We're seeking a Senior Software Engineer to build the AI agents patients actually talk to. Our agents hold real conversations with real patients every day, and each one has to be accurate, safe, fast, and kind. In this role, you'll own agents end to end: conversation design, prompts, tool calling, and evals, plus the APIs and data that make an agent work in production. This is a full-stack AI role, not a research-only seat. You'll take an idea from concept to an agent in front of patients quickly, then use live conversation data to make it better. In your first three months, you'll ship multiple production agents with our health system partners, build the eval suite every agent change must pass before it ships, and measurably improve how our agents perform in production.

You'll work directly with our SVP of Engineering and engineering leads, alongside product, clinical, and customer teams, on a small team where your work reaches production fast. The ideal candidate is a hands-on builder and fast learner who is comfortable with ambiguity, communicates clearly with engineers, nurses, and health system executives alike, and is motivated by making healthcare work better for patients.

Responsibilities

• Design, build, and ship AI agents end to end, including conversation flows, prompts, tool and function calling, retrieval, memory, and the backend services agents depend on.

• Build evaluation systems (offline test suites, simulated conversations, human review loops, and production monitoring) so we know an agent is good before and after it ships.

• Improve instruction following, tool-calling reliability, latency, and cost across our LLM and speech stack.

• Help design and build our next-generation voice agent platform, choosing the right models and architecture for each task based on quality, latency, and cost.

• Partner with clinical, product, and customer teams to turn workflows like scheduling, refills, billing, and post-discharge check-ins into reliable agent behavior.

• Dig into real conversations and failure cases, and close the loop with prompt, tool, model, or product changes.

• Build internal tools to observe, monitor, and analyze agent behavior at scale to identify common failure modes and patterns.

• Track what's new in models and tooling, and bring the useful parts into production quickly.
  

Qualifications

• 5+ years of professional software engineering experience, including 1+ year building and shipping LLM-powered applications in production.

• Strong proficiency in Python and/or TypeScript, comfortable across backend APIs and data, with enough frontend to ship a complete feature.

• Hands-on experience with prompting, tool use and function calling, RAG, structured outputs, and agent orchestration.

• Experience designing evals and using data to decide what to fix next.

• Product sense: you care whether the patient got their appointment booked, not just whether the model metric moved.

• Clear written and verbal communication with technical and non-technical teammates.
  

Preferred

• Voice AI or real-time speech experience, including STT, TTS, voice activity detection and turn-taking, and telephony (Twilio, SIP, WebRTC).

• Fine-tuning or post-training experience with open or hosted models.

• Experience in healthcare or another regulated environment, including HIPAA, EHR, and FHIR.

• Experience with workflow orchestration (e.g., Temporal).

• Familiarity with LLM observability and tracing tools.
  

Salary and Benefits

We offer a competitive salary and meaningful equity, along with comprehensive health, dental, and vision insurance, a 401(k) plan, and [learning budget / conferences / additional benefits].

The typical salary range for this role is $160,000 to $210,000 USD, depending on skills, qualifications, and relevant experience.