Hey Soham,
That’s one of the most stacked inros I’ve seen here from GenAI to Solana to real-time infra with AWS Fargate, you’ve touched some pretty serious ground. RepoGPT in particular sounds like a powerful devtool 95% retrieval accuracy with RAG is actually asupr impressive!
Quick q — how did you handle embedding/document chunking in RepoGPT with LangChain + Gemini? Been tinkering with similar setups myself and would love to know your approach.
Also, if you’ve got any AI startup ideas or side projects you’re exploring, definitely drop them in our AI Ideathon thread some of the best ones will get support from the team here. Would love to see what you’d bring to the table!
Super excited to have you here, man. Let’s build 