Case study
Sentinel - AI-Powered Engineering Operations Platform
Our in-house operations platform: plans, releases, costs and CRM in one place, with AI coding agents that work under human approval.
Overview
Running engineering across several companies means plans, repositories, releases, infrastructure costs and client conversations end up scattered across a dozen tools. Sentinel is the internal platform we built to run it all from one place — and to put AI agents to work safely. Its guiding rule: AI drafts and proposes; a person accepts.
Key features
- Boards and portfolio milestones with delivery metrics
- AI coding agents (Claude Code, Codex) that open pull requests only after a manager approves
- "Code Brain" repository maps powered by our open-source cgb
- Release pipelines over SSH, WinRM, S3 and AWS ECS with approvals and rollback
- Asset, cost and renewal tracking — domain and TLS expiry, plus a compliance calendar
- CRM and prospecting with AI-drafted outreach, a shared mail workspace and a hosted MCP server
Tech stack
- NestJS 10 and Prisma on PostgreSQL 16
- Redis and BullMQ job queues, S3-compatible storage
- React 19 and Vite web app with Clerk authentication
- Electron desktop runner for macOS and Windows
- Docker Compose, pnpm monorepo
Status
In daily internal use at vxlabs. Want something similar for your team? Talk to us.