00.01 PRODUCTION AI

Systems that
think with you.

We build production AI systems — RAG pipelines, MCP servers and agentic search products — integrated with the rigour of 20+ years of engineering practice.

Start a project ↗ See the BookStack case study
11
weeks to prod
871+
tests shipped
90%
coverage
18y
delivery
00.02 · LIVE SCHEMATIC
bookstack rag · hybrid search
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
01.00 WHAT WE BUILD

Three surfaces, one discipline.

Every engagement spans ingestion, retrieval and response. We design each surface with the same engineering rigour — TDD, IaC, observability, docs.

01

Ingestion

Webhook-driven pipelines that keep your knowledge base in sync with vector indexes in near-real-time.

02

Retrieval

Hybrid dense + sparse search with RRF fusion and cross-encoder semantic reranking.

03

Response

Answers generated with hierarchical citations, traceable back to Shelf → Book → Chapter → Page.

02.00 FEATURED SYSTEM

BookStack RAG — live schematic.

A production-grade RAG system on GCP. Trace the signal as it travels from a user question to a cited answer.

02.01 · FULL PIPELINE
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
11
weeks to production
194
commits
871+
tests
90%+
code coverage
03.00 DELIVERY

AI-augmented SDLC.

03.01 · HOW CODE GETS WRITTEN HERE
prd → blueprint → codegen → ci → prod
META-PROMPT SPEC GREEN DEPLOY PRD INTENT 01 Blueprint LLM 02 Docs LLM 03 Codegen TESTS FIRST CI/CD GH ACTIONS Production OBSERVED
// 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 →