About

Hey, I’m Ashwin. I’m a software developer based in Bengaluru, India.

I spend most of my time building web applications and practical AI workflows. Right now I’m at Doozie Software Solutions, working on LLM and full-stack problems in real product work. Before that, I interned across a couple of teams, built real-time apps with WebSockets, worked with React, and learned by shipping.

This site is where I collect what I’m learning, what I’m testing, and the occasional opinion about developer tooling.

What I’m focused on

• Full-stack product work with AI in the loop.
• Real-time speech and audio intelligence, which I work on in my own time in a small personal lab. The main project is MendSpeech, a real-time dictation system covering streaming ASR, CTC decoding, confidence calibration, and latency measurement.
• AI safety, model evaluation, and tooling with the AI Risk Evaluation Workbench.
• Writing more of the useful things I learn while building.

Writing

I also write on Hashnode from time to time, but this blog is where I want the longer-form thinking to live.

Elsewhere

You can find most of my work on GitHub. The best way to reach me is X — send me a DM and I’ll reply. You can also find me on LinkedIn.