← All work

Founding Engineer & Technical Lead

Vexor

An AI-powered legal case-management platform that converts incoming notifications into structured triage decisions.

Initial delivery
2-3 months
Team growth
+2 engineers
Integration
Galaxy CRM

01 · Context

The system behind the product

Took the product from concept to a stable production system for a European legal firm group, then expanded the engineering team.

02 · Architecture

How the system is shaped

  1. 01

    Incoming legal notifications enter a Python processing pipeline that extracts structured case information and applies scenario-aware triage logic.

  2. 02

    The AI decision layer publishes comments and triage outcomes into Galaxy CRM so legal teams stay inside their existing operating workflow.

  3. 03

    Azure Container Apps and private endpoints provide an isolated production boundary for application and integration services.

03 · Decisions

Trade-offs made explicit

Integrate where work already happens

AI output is surfaced inside Galaxy CRM instead of forcing legal users into a second disconnected interface.

Structure before automation

Notification content is normalized into explicit case data before scenario rules and LLM-assisted decisions are applied.

Scale the team after the spine worked

The initial production system established the core boundaries before two additional Python engineers were onboarded.

04 · Outcome

What changed

Delivered the initial production platform in 2-3 months and integrated its AI decisions directly into the client's existing CRM workflow.

  • Moved from concept to a stable production system in 2-3 months.
  • Reduced workflow fragmentation by placing generated case comments and triage decisions directly in the client's CRM.
  • Expanded delivery capacity with two additional Python developers while retaining architectural and client ownership.

05 · Capabilities

What the work involved

  • Legal document processing
  • AI decision pipelines
  • CRM integration
  • Team formation