Case Study — AI-First Engineering Team

Building an
AI-First Team

A case study in AI-augmented software delivery: 5 engineers modernizing a legacy .NET monolith into microservices, with an Agentic RAG system that lets AI query 10 years of system history.

5
Engineers
9 mo
Timeline
AI Multiplier
~44 MM
Effective Capacity
5
Services Shipped
Zero
Downtime

System Architecture

Strangler Fig pattern — incremental migration via YARP gateway

View details →

Migration Timeline

4 phases over 9 months — Phase 0 invests in AI foundation

All phases →

Services

5 services to extract + Payment (frozen in monolith)

All services →

AI Knowledge System

How we built an Agentic RAG pipeline so AI can query 10 years of legacy history

Try it →

1. RAG Pipeline — Building the Knowledge Base

Generate

Python scripts generate realistic 10-year artifacts: 13K+ JIRA tickets, 667 incidents, 500+ deployments, team events, code, configs, meeting notes

Embed

All artifacts chunked and embedded using all-MiniLM-L6-v2 (local, 384 dims, no API needed) into ChromaDB vector store

Store

21,127 documents in ChromaDB (311 MB) — searchable by semantic similarity, doc type, module, era, year

2. Agentic RAG — AI Decides What to Search

Naive

Traditional RAG: embed question → find similar docs → paste into prompt. Works but no reasoning about WHAT to search

Agentic

LLM receives 13 tools as function declarations. It decides which tools to call, chains multiple searches, then synthesizes

Loop

Multi-step: LLM can call tools → read results → call MORE tools → synthesize. Up to 5 iterations of reasoning

Two Access Patterns — Same RAG Backend

Developer Flow (Copilot + MCP)

Developer asks in VS Code Chat

→ Copilot analyzes question

→ calls MCP tools (stdio JSON-RPC)

→ ChromaDB returns docs

→ Copilot synthesizes answer

✓ Zero API cost — Copilot is the LLM

✓ 13 tools via MCP server

Web Flow (Gemini + Function Calling)

User asks on Ask AI page

→ Gemini receives 13 tool definitions

→ decides which tools to call

→ Python API → ChromaDB returns docs

→ Gemini synthesizes answer

✓ Same tools, different LLM

✓ Deployable as web app

13 RAG Tools Available

search_legacy_history
search_by_type
search_by_module
search_timeline
search_combined
search_code
get_system_overview
get_incident_stats
get_code_inventory
get_team_timeline
get_module_health
find_related_artifacts
list_available_filters

Risk Summary

Details →
F1CDC Data InconsistencyMediumHigh
F2Legacy Payment OutageMediumHigh
F3AI Business Logic ErrorsHighHigh
F4Cascading Canary FailureLowHigh
F5Key Engineer DepartureMediumMedium

Key Milestones

Timeline →
M1
AI Toolchain Deployed
Phase 0
M1
CI/CD + IaC Ready
Phase 0
M1
YARP Gateway Live
Phase 0
M1
Comms Pilot (Staging)
Phase 0
M2
Payment ACL Operational
Phase 1
M3
Travel Booking Go-Live
Phase 1
M4
Event Management Go-Live
Phase 1
M6
Workforce Go-Live
Phase 2
+5 more milestones →