Sekharendu Dey

Hey, welcome ^_^. I'm a software engineer from India, working my way into AI engineering. These days, I'm focused on LLM backends, retrieval, and the infrastructure that keeps them fast and reliable.

I document my journey publicly on X, sharing the wins, the dead ends, and everything in between. I like picking problems that feel a bit out of my league. Going from completely stuck to finally figuring it out is my favourite part.

What I've been building

• LocalCortex: a fully local RAG pipeline with adaptive chunking and retrieval evaluation (Recall@K, MRR).
• StreamMind: a single backend layer for LLMs that handles streaming, RAG, agents, memory, observability, and fallback across providers.
• AI Adapter: drop-in telemetry for AI agents that captures latency, token usage, and errors with no code changes.

Away from the keyboard

You'll usually find me playing cricket, working out at the gym, or going for a run. I love anime for stories that pull me into another world, and non-fiction books that make me see this one a little differently.

And some days, after all the building, learning, and overthinking, I just sit back, watch the clouds drift by, and do absolutely nothing for a while.

Just a guy who refuses to give up on his dreams.