Senior Product Engineer

Talkie builds AI Agents for healthcare. Our agents handle patient-clinic communication end-to-end over voice calls, web chat, and SMS, so patients get 24/7 access to care and busy practices never miss a conversation. Every month, our AI Agents handle close to a million real patient conversations across the US and Poland. Our customers include primary care and specialty practices, and hospitals.

We're looking for a Senior Product Engineer to join our Platform team. You'll build the features that power our AI Agents, in the backend and the UI. Much of that work touches telephony, clinical systems such as EHRs, and LLMs.

You'll own features until they run in production. You research the problem and shape requirements with our CPO. You propose the interaction and UI, design the architecture, build it across the stack, ship it, and watch how it behaves with real patients on real calls. You'll work alongside other Engineers, take requests from our Customer Success and Conversation Design teams, and work with many people across the product organisation.

Our engineers work with many AI agents at once. One agent implements a feature, another reviews it, and a third digs through logs or researches an unfamiliar API. This role requires fluency in that way of working.

• →

  Own product features end-to-end. Research the problem, refine requirements with the CPO, propose the user interaction, and design a clean, usable UI. Then build it, ship it, and monitor it.

• →

  Build across the stack. You'll write backend services and APIs in TypeScript and Node.js, frontend in React, and AWS infrastructure as code.

• →

  Integrate LLMs into the product. Design how our agents use models and tools. Build evals that show whether a change made things better, and trade off quality, latency, and cost on purpose.

• →

  Integrate with outside systems. Our agents depend on SIP telephony and call routing, EHR systems, LLM providers, and other services.

• →

  Shape technical and architectural decisions. Take part in the team's design discussions, write and review technical proposals, and help set the direction of the platform. Keep it easy to change and maintain, and make sure it handles real-time calls efficiently.

• →

  Mentor other engineers. Review their designs and code, pair with them on hard problems, and help them grow, including in how they work with agents.

• →

  Turn open questions into plans. When a new problem or technology comes up, form a hypothesis, build a prototype, and measure it. Then come back with a list of trade-offs and a proposal the team can build on.

• →

  Keep production healthy. Monitor what you ship, cut alert noise, and find and fix performance bottlenecks.

This is how we work every day, and we expect you to help us improve it.

• Orchestrate agents. You run several coding agents in parallel for implementation, code review, log and incident investigation, and research. You know how to direct them, check their output, and correct it.

• Improve the codebase for agents. You know what makes a codebase easy for agents to work in, such as clear module boundaries, good agent instructions in AGENTS.md or CLAUDE.md, and fast feedback loops. You make ours better.

• Build skills and agent workflows. You write and maintain reusable skills, commands, and workflows that encode how the team works, so agents get it right the first time.

• Share what works. You teach the team effective agent workflows, and you know when not to trust the output.

We care less about years of experience and more about what you've shipped, owned, and learned.

• Product ownership. You've shipped and owned a product or major features in production, end-to-end. When the path isn't clear, you find one.

• Strong architectural experience. You build software that stays easy to change and maintain as it grows, and you can explain the trade-offs behind your decisions.

• Fullstack TypeScript. You have solid hands-on experience with Node.js on the backend, React on the frontend, and MongoDB or another database.

• Cloud and infrastructure as code. You have hands-on experience with AWS or another major cloud provider, and with IaC tools. We use Terraform and Pulumi.

• AI-native development. Coding agents are your main way of working, as described above, and you can show us how you use them.

• LLM integration in production. You've built LLM-powered features with evals, and made informed latency and cost trade-offs.

• Healthcare compliance awareness. You understand what building software under SOC2 and HIPAA means in practice, and how it shapes decisions about data, access, and logging.

• Product sense for UI. You can propose a user interaction and design a clean UI without a detailed mockup.

• English. You speak and write English at C1 level or above. Most of our written work is in English, and our customers are in the US.

• Telecom. You've worked with SIP, call routing, or VoIP. If you haven't, that's fine. When you hear "SIP integration", we want your reaction to be "I'll learn it", followed by a plan for a prototype.

• EHR or other healthcare integrations. You've worked with clinical systems or other regulated domains with many integrations.

• You want detailed specs before you start.

• You see coding agents as autocomplete, or as a fad.

• You only want to work on one side of the stack.

• "I've never done SIP" ends the conversation for you instead of starting a prototype.

• You find it hard to judge when to debate a decision and when to build a quick prototype and measure it.

• Learn how a patient call or chat travels through our platform end-to-end, across telephony, the AI Agent, LLMs, and the EHR.

• Ship your first features end-to-end, from research through review with the CPO to production.

• Take on production health. Reduce the number of alerts in our Slack monitoring channel and suggest performance improvements.

• Make improvements to our codebase or agent workflows that the rest of the team adopts.

• Understand the core parts of the platform in depth and drive feature work across them.

• Lead architectural decisions that improve the platform and make sure it handles real-time calls efficiently.

• Mentor other engineers on the team, so your knowledge of the platform and of agent workflows spreads beyond you.

• Become the person the team goes to for turning troubleshooting findings and internal issues into clear implementation plans. Those plans come with prototypes, trade-off lists, and walking skeletons that deliver value and that we can build on.

We are friendly, direct, curious, and ambitious. We value a growth mindset and treat failure as a chance to learn. Asking questions is the standard here, and we expect thorough investigation. When something's broken, we say so, propose a fix, and fix it. We also like to have fun. Dancing till the early hours, karaoke nights, board games, and team sports are a long-standing tradition at Talkie.

We want our recruitment process to be respectful and collaborative, so you can be yourself and get to know us. The whole process usually takes 2 to 4 weeks.

1. Intro call. A short intro of 10 to 15 minutes.

2. First interview. A 30-minute online conversation about mutual fit.

3. Small hands-on task. We may send you a short task (up to 1 hour of work) to do on your own, using your usual agent setup. We'll talk it through in the second interview, including how you worked with LLMs and agents along the way. 

4. Second interview. A 60 to 90-minute online session with hands-on work and a system design discussion. Have your dev environment and agent setup ready, because we want to see how you work.

5. Reference check. We'll ask for references from people you've worked with.

6. Offer. If all goes well, we'll make you an offer. If not, we'll share detailed feedback based on the interviews and references.

We want to hire as soon as we find the right person, so apply now.