FluidLabs

We're building an AI-powered document processing platform that automates complex back-office workflows — from intake and validation to approval and ERP synchronisation. The system is already in production and handles real financial flows; you'll be joining at the scaling stage.

What you'll work on

• AI orchestration layer: LLM tool-calling pipelines (Claude / OpenAI), MCP server (4 domain tools), context and retry management
• Integrations: Microsoft 365 Graph API, DocuSign REST API, Azure Document Intelligence
• Validation & routing engine: business rules, confidence scoring, exception handling
• Backend API: REST + SSE, OAuth2/OIDC, audit logs, OpenTelemetry
• Frontend: React 18 + TypeScript + Tailwind dashboard (agent status, approval queue, analytics)
• Infrastructure: Docker, GitHub Actions CI/CD, Azure / AWS cloud

Required

• Node.js + TypeScript — 4+ years in production, architecture-level decisions
• LLM API integration — tool calling, prompt design, context management (OpenAI, Anthropic, or Gemini)
• REST API design + OAuth2/OIDC implementation
• PostgreSQL — schema design, query optimisation, migrations
• Microsoft 365 Graph API or similar enterprise API (SharePoint, OneDrive, Teams)
• Azure Document Intelligence / AWS Textract / similar document AI service
• React 18 + TypeScript — you can own the frontend when needed
• Docker + CI/CD (GitHub Actions or equivalent)
• Cloud deployment (Azure or AWS)
   

Nice to have

• DocuSign API experience
• MCP (Model Context Protocol) — building or consuming MCP servers
• pgvector / vector similarity for deduplication
• OpenTelemetry — metrics, traces, structured logs
• Tailwind CSS + shadcn/ui
   

We're looking for

Someone who has shipped AI-integrated features in production — not just experimented with them. You take ownership end-to-end: design → implementation → deployment → monitoring. You can make pragmatic architecture calls and explain trade-offs clearly.
 

Senior bar: you've led technical design of at least one non-trivial integration (enterprise API, document pipeline, or payment flow), and you don't need to be told what to research.
 

Offer

• Fully remote, async-friendly, EU time zone overlap (11:00–19:00 Kyiv)
• Small product team no bureaucracy, direct access to founders
• 24 days paid leave + public holidays
• 3-month probation period
   

How to apply

Apply via Djinni or send your CV + a short note about your LLM integration experience. Тell us: what was the most complex API integration you