Software engineer for AI voice agents

We are hiring a mid-level full-stack engineer to build and run our AI voice and chat agent platform: Ruby on Rails for the core backend, Python for AI orchestration, and TypeScript for the frontend and microservices, all on Google Cloud. You own your features end to end: backend, frontend and whatever infrastructure changes it takes to get them live, then you run them in production.

• Build new features in our voice AI product across the stack: Rails backend, Python AI orchestration, TypeScript frontend and microservices, plus the infrastructure changes they need.
• Handle incoming customer requests and maintain the voice AI features already in production.
• Build and extend our public REST APIs and MCP server API-first: OpenAPI spec, then code, then contract tests, without breaking existing clients.
• Integrate third-party providers (calendars, booking, ticketing, telephony, CRMs) as tools our AI agents call during live phone calls and chats.
• Work on the voice side of the product: real-time call rooms on LiveKit, speech-to-text and LLM providers, call results and transcripts.
• Write background jobs for imports, outbound calls, reporting and cleanup.
• Write tests for what you ship (Minitest, RSpec, Capybara system tests, Jest) and keep CI green.
• Own your changes in production: watch logs and alerts on Google Cloud, track errors in PostHog, fix what breaks.
• Review teammates' pull requests and keep the docs current.

• 3+ years building and shipping web applications professionally, in any language.
• Strong programming fundamentals: data structures, clean code, debugging, and reasoning about performance.
• Backend experience
• Modern frontend with a component framework, ideally React with TypeScript. Ours is React 19 and TypeScript 5.
• SQL and relational databases: schema design, indexes, reading a slow query and fixing it. We use MySQL.
• API design, error envelopes, versioning, backwards compatibility.
• Automated testing on both backend and frontend, and the habit of writing the test with the change.
• Git and pull-request workflow, with CI you are expected to keep passing.
• Docker for local development.
• Comfortable reading logs and debugging a live issue, not only local bugs.
• Hands-on agentic coding with Claude Code, Codex or other harnesses: prompting, giving the agent the right context, and reviewing and owning what it produces.
• Clear written English for PRs, specs and docs.
• Будет плюсом: Voice and real-time: LiveKit, WebRTC, telephony/SIP (Twilio or similar), speech-to-text such as AssemblyAI; LLM applications: OpenAI API, function calling / tool use, LangChain, RAG with vector databases (Weaviate, Pinecone); Frontend tooling: Tailwind CSS v4, shadcn/ui and Radix, Webpack via Shakapacker, react-on-rails server-side rendering, Storybook, Zod; Ruby on Rails; Google Cloud: Cloud Run, Cloud Logging and alerting, Cloud Storage; GitHub Actions for CI/CD; API contracts: OpenAPI 3, Redocly linting, contract tests; Product analytics and errors: PostHog; Security basics: Brakeman, bundler-audit, OAuth flows, multi-tenant data isolation; i18n: shipping UI in several languages (we ship English, German and Spanish); Python: to contribute to our voice-agent worker (preferred, not required)

• Условий в вакансии нет