At Edvanta Technologies, I was a core engineer on ThoughtJumper, a knowledge-graph learning platform. I built its Vue.js workspace around an interactive AntV G6 canvas and redesigned the MongoDB graph model so editing one node no longer rewrote the entire graph.
I cut API response times by 90% and Redis cache footprint by 98% by compressing graph payloads before caching. I also implemented a Graph-RAG pipeline that gives each user a separate vector index, and developed real-time collaboration and a Chrome extension for capturing content into a graph.
For a US client, I delivered an end-to-end SaaS product with Next.js, Supabase, Drizzle ORM, and Stripe, including multi-tenant security with PostgreSQL Row Level Security. My projects include Kreation, an AI thought-capture SaaS, and event-driven RAG backends using FastAPI, AWS SQS, and Pinecone.
