The opportunity
Our backend tech stack consists primarily of Python Flask web apps with our TaskTiger scheduler handling many of the backend asynchronous task processing chores. Our data stores include MongoDB, PostgreSQL, Elasticsearch, and Redis.
What you'll do
Agents. Build Close’s agentic platform and customer-facing AI experiences.: Voice Agents, Custom Agents, and Ask Chloe are already in flight. You’d work on the shared intelligence and orchestration layer that powers AI-driven processes across the product. That includes the ecosystem around our APIs and MCP surface, which enables external agents like Claude and ChatGPT to operate Close. This team is pushing deeper into applied AI, retrieval, and agent infrastructure more than ever before.
Communications. Own the voice, SMS, and email infrastructure powering Close —: Twilio, WebRTC, WebSockets, AssemblyAI, ElevenLabs, and more. You’ll help build our voice, messaging, and conversational AI agents while scaling real-time AI across calls, email, and calendar sync infrastructure.
CRM. Build the structured context layer that humans and agents both depend: on. We're rebuilding the data model so it flexes with real businesses, making the CRM agent-ready (retrieval, traversal, action coverage through MCP and our public API), and shipping the next generation of AI-native CRM features (AI Enrich, Autofill, AI Search). Operating on billions of Mongo documents with Elasticsearch underneath.
Growth. Run the experiments and own the billing infrastructure that turn: trials into paid customers. Stripe metered billing, AI credit top-ups, the activation and conversion funnel, and the team's biggest bet right now: agentic onboarding (an agent that walks new customers through setup, configuration, and first comms). Hypothesis-driven, metric-first.
A seasoned engineer. Python is our backbone, but perhaps you’ve worked across: Go, Rust, or TypeScript. You pick the right tool for the problem rather than retreating to what you know. You’ve seen a variety of problems, can collaborate and self-direct.
Building AI. You've shipped LLM-backed features for real users and have a POV: on where they're trustworthy, where they fall over, and how to get them production-ready when customers bet revenue on the output. We use Pydantic, Temporal, LangFuse, and other modern AI infrastructure tooling to power our agent platform.
What they're looking for
- Working with AI in your day-to-day. You use AI tools in your own workflow to: ship faster, write tighter code, and reason about unfamiliar parts of the codebase. We fund the use of AI tools (Claude Code, Codex and other best-in-class developer tools) and treat learning and experimentation as part of the work.
- A builder first. You’d rather get the first iteration of a feature in front: of fifty customers than spend three weeks fretting over a perfect PR. You ship.
- Opinionated about API design for modern software. You've shipped: internet-facing APIs and you think about who's on the other end - apps, agents, humans poking around in docs - and what each of them needs.
- Battle-tested. You've debugged incidents where latency budgets didn't hold,: owned a system everyone else relied on, or carried a pager for something with real customer impact.
