About

Hi, I’m Tushar.

I lead engineering at Littlebird where we're building personal AI that fiction promised. A thought partner that understands your world and extends how you think..

Getting this right sits at the intersection of a bunch of hard, mostly unsolved problems - how do you capture someone's world, keep learning from it, and then actually do useful things with that understanding? That breaks down into context capture, retrieval, memory, knowledge graphs, agent orchestration, SLMs, integrations, intuitive interfaces that make the complexity disappear and more - but none of these work in isolation. I think across all of it. Personal AGI is a systems problem, and the magic is in the compounding.

Past

I was previously the technical cofounder at Wokay unifying different aspects of work collaboration - chat, tasks, docs into one app where agents work alongside humans.

Before that, I was at 100ms(posts) where I worked on core infrastructure for live video - WebRTC, HLS streaming & recording. I also led the Web team to build the Web SDK and the AI team to build features like transcription + summarization, noise suppression and virtual background.

Before 100ms, I was working at Goldman Sachs in platform engineering, which is a polite way of saying I built internal tools that other teams couldn't avoid using. Before Goldman, I interned at IDfy working on training models for problems around object detection, OCR and liveness checks.

I studied electrical and electronics engineering at BITS Pilani. I've always had a thing for hardware - the way a well-made physical object earns a kind of trust that software never quite does. At some point I want to build personal computing devices that feel like that.