Proof, not promises
VertexMind.ai is a new practice — so instead of a wall of client logos, here's the actual code. These are production-style microservices built and open-sourced, plus the architecture patterns we bring to client work.
RAG & agentic microservices, on GitHub
Real, inspectable repositories — not screenshots.
docchunker-api
Document chunking service — splits heterogeneous source documents into retrieval-ready segments while preserving structural context.
retrieval-service
Vector retrieval microservice — the query-side layer that turns a chunked knowledge base into ranked, relevant results.
reranker-service
Cross-encoder reranking service that sits after retrieval to push the truly relevant results to the top — the difference between a demo and something people trust.
schema-mapping-studio
A FastAPI production microservice that uses an LLM to propose JSONata schema mappings between systems — CRM/contact-center signal normalization across Zendesk, Genesys, and Salesforce — gated behind a human-approval step.
What we apply in client engagements
Categories of system design pulled from enterprise AIOps and network-operations work — described here at the pattern level. Named case studies are available on request.
Policy & governance layer for autonomous agents
A shared policy-enforcement service that sits in front of any autonomous agent — evaluating proposed actions against configurable rules before they execute, so agents get more autonomy without losing an approval trail.
Decision dashboards for cloud & network operations
ML-driven views that convert raw telemetry and cost data into a short list of decisions worth making — built to reduce dashboard fatigue, not add to it.