RCA Insights for a medical-device company
Local-LLM embeddings over 400+ engineering postmortems so root causes can be found and closed.
An internal tool at a medical-device company that embeds Jira postmortems with a local LLM and makes them searchable by root cause, alongside reusable AI-assisted testing workflows for the engineering team.
More than 400 postmortems held lessons nobody could find. Related failures were investigated from scratch, and test writing was a steady drag on delivery.
Built RCA Insights in Next.js and TypeScript with local embeddings so the data never leaves the company. Related incidents and their root causes surface together. Separately, set up AI-assisted testing workflows the team reuses, and shipped accessible clinical visualizations with five-language and right-to-left support.
More in AI & Automation
Blog Factory for an agency client
Research, drafting, images, review and CMS publishing for a marketing agency, in one system non-engineers can run.
Client buildSales audit deck
Prospect-site audit that produces a branded PDF deck sales can walk through on a first call.
Client buildAI visibility checker
Snapshots of how a business appears across Google AI Overviews, ChatGPT Search, Gemini and Perplexity.