JD Addy
AI Agent Developer | Voice AI, Neo4j RAG & Document AI
A document AI platform that reviews candidate documents during hiring. The startup's existing codebase was incomplete and not working, so we rebuilt it into a production system that reads, sorts and checks documents automatically. It now processes 1.5M documents a month, cutting a two-week review to under three minutes, for clients including McDonald's and government agencies.
0
19
A PDF inspection reports app for home inspectors. Inspectors filled long wind-mitigation checklists by hand, then rebuilt them as reports back at the office. The web app turns the official checklist into a guided, step-by-step form on site, with photos attached as evidence; reports go to the client as a share link or branded PDF, and every client and report lives on one dashboard. 62 inspection reports have run through it so far.
0
24
GoalPost, an AI-searchable community knowledge graph. What a community knows about its people, goals and progress was scattered across notes, chats and documents. In GoalPost, people, goals, updates and documents are nodes in one living graph; uploaded documents are read and their key points extracted automatically, and the built-in AI answers questions from the graph, then redraws the canvas around its answer.
0
29
Smith, an autonomous AI developer agent. It picks up Jira tickets, writes the fix, runs the checks and opens a pull request with an evidence report: the bug, what changed and proof it is fixed. A reviewer's comment sends it back to rework the branch, and it keeps what it learns in a 6,000+ node knowledge graph. So far: 129 pull requests across 26 projects, 101 merged without changes.
0
34