Feature-Request Scoring
An LLM scores the open feature requests in a GitHub org on five dimensions and explains each score, so a product manager can rank, sort and filter the backlog. Shown here on Cloudinary's public repos.
LiveAI Solutions Builder
Hi 👋
I build AI solutions around what customers need, and then keep them running.
With years of experience in building and supporting products, as well as optimizing workflows, I now build with all 3 in mind.
Above all, I believe in learning by building - and giving back to the open-source tools that I use. This site showcases some of my recent personal projects.
I'm available for freelance projects and open to full-time roles.
Building LLM reasoning into products and workflows directly through a model API (e.g. Anthropic API), no framework.
An LLM scores the open feature requests in a GitHub org on five dimensions and explains each score, so a product manager can rank, sort and filter the backlog. Shown here on Cloudinary's public repos.
LiveAn AI agent redistributes a payment schedule from plain-language chat instructions, with every proposed number independently verified before it's shown.
LiveAn LLM reads a photo or PDF of an Israeli invoice and extracts the supplier, line items, VAT and totals, then separate code checks every number before the document is approved.
LiveOrchestrating business processes with AI agents and human-in-the-loop reviews, e.g. Camunda/BPMN.
A BPMN-driven workflow that triages incoming documentation feedback, with a human-in-the-loop review step before anything is filed.
LiveA documentation fix contributed and merged into Camunda's public docs repo.
Live View PRBuilding custom agents with frameworks like LangChain: chains, tools, and memory.
Four documentation fixes contributed and merged into LangChain's official docs repo.
Live View PRsA hands-on introduction to building agents with LangChain.
In Progress