DocuMind AI
Production-grade RAG is usually treated as either a wrapper over ChatGPT (no isolation, no cost control, no observability) or as enterprise infrastructure (Pine…
View case study
I design systems.
I write the code.
I ship the product.
Fullstack engineer with 4 years designing and delivering software systems end to end — SaaS platforms, data-intensive applications, and AI-powered automation pipelines.
I think at the architecture level first: understand the business constraint, pick the right tools, decide on tradeoffs. Then I execute — production code, integrations, testing, shipping.
Systems I've designed
and built.
Production-grade RAG is usually treated as either a wrapper over ChatGPT (no isolation, no cost control, no observability) or as enterprise infrastructure (Pine…
View case study
Job searching is a high-volume, repetitive process that still requires significant manual effort per application — reading the posting, tailoring the cover lett…
View case study
Existing QR management tools were either too limited for enterprise workflows or too complex for small businesses. YRQR needed a platform that could serve both …
View case study
Farmers and agricultural technicians had no digital tool to model the nitrogen footprint and carbon credit impact of their practices. Decisions about soil manag…
View case study
EV charging fleet operators had no unified interface to monitor station status, track usage across locations, and manage charging schedules. Operators were mana…
View case study
Transport and logistics companies face structural pressure to digitalize and reduce environmental footprint as part of EU policy goals — but lack the tooling an…
View case study
Code I've shipped
publicly.
Multi-tenant RAG SaaS — PDFs in, cited answers out. Hybrid search, streaming citations, async ingestion, per-tenant cost tracking from day one.
Marketing-oriented MCP server on top of trivago's official MCP — aggregated market intelligence inside Claude, Cursor, VS Code, or any MCP client.