02.00 PROJECTS

Selected work.

Production systems we designed, built and deployed — architecture, cadence and outcomes.

02.01 CASE STUDY

BookStack RAG

A production-grade Retrieval-Augmented Generation system on GCP that turns a BookStack knowledge base into an intelligent, searchable system with AI-powered answers and source citations.

CLIENT · EQUAL EXPERTS
TIMELINE · 11 WEEKS
STATUS · LIVE
02.02 · ARCHITECTURE
ingest · embed · search · rerank · answer
QUERY INGEST UPSERT / SEARCH CANDIDATES CONTEXT CITATIONS User SLACK / MCP 01 BookStack WEBHOOK 02 Embeddings DENSE + SPLADE 03 Vector Index HYBRID 04 Reranker VERTEX AI 05 Gemini 2.5 ANSWER
—

Hybrid vector search

Parallel dense and sparse embedding retrieval, combined with Reciprocal Rank Fusion for comprehensive relevance.

  • → Dense embeddings via Vertex AI text-embedding-005
  • → Sparse embeddings via FastEmbed SPLADE
  • → RRF fusion combining both retrieval paths
—

Semantic reranking & RAG

Answer generation with hierarchical source citations from reranked search results.

  • → Vertex AI Ranking API for semantic reranking
  • → Gemini 2.5 Flash for answer generation
  • → Hierarchical citations (Shelf › Book › Chapter › Page)
—

MCP Server

Model Context Protocol server exposing the knowledge base as a tool for AI agents.

  • → Streamable HTTP transport
  • → Claude Desktop & Claude Code integration
  • → Direct search pipeline access
02.03 STACK
AI / ML
Vertex AIGemini 2.5 FlashFastEmbed SPLADERanking APIClaude Haiku
Infrastructure
GCPCloud RunCloud TasksFirestoreTerraform
Development
PythonFastAPIPydantic AIpytestGitHub Actions
Integration
MCPSlack Socket ModeBookStack APIBlock Kit UI
02.04 ALSO DEPLOYED

Other systems.

Belgium

Boekhoudkantoor Vervaet

AI chatbot with RAG on AWS Bedrock Knowledge Base.

AWS BedrockRAGPythonLambda
UK

Travelopia

Scalable lead-allocation engine with AI-augmented delivery.

Node.jsNestJSAWSGenAI
USA

SiriusXM

Identity management with passkey support for streaming.

ScalaCatsDynamoDBAWS CDK
// NEXT STEP

Ready to build a system that thinks with you?

An 8-week engagement takes you from question to deployed answer.

Start a project →