All work
2024 – PresentTech Lead → Project ManagerClosed-source
REST Solution
A grounded chatbot over an indexed knowledge base, paired with a serverless pipeline that turns AI-generated Markdown into client-ready Word documents in Azure Blob.
Azure Cognitive SearchAzure OpenAIAzure Blob StorageDjangoAzure Functions
2
Tiers
Azure
Cloud
2–3
Team size
Summary
Two-tier AI tooling for a US-based client's internal team. A Django chatbot retrieves the right material from an Azure Cognitive Search index and asks Azure OpenAI to ground its answer; a separate Azure Function takes Markdown content, converts it to .docx via python-docx, and writes to a fixed Azure Blob container with filename-idempotent uploads. Markdown is the contract between the tiers — debuggable in a text editor; the Blob container is treated as a current-state surface, so re-runs never pollute the deliverable.
Highlights
- Hybrid keyword + semantic retrieval on Azure Cognitive Search
- Markdown → DOCX pipeline as a separate Azure Function — independent scaling
- Filename-idempotent Blob uploads — re-runs never pile up duplicates
This project is closed-source (built for a Kcube AI client). I'm happy to walk through the architecture, trade-offs, and code on request.