I am from Chennai, and a lot of what drives me still comes from there: loyalty, curiosity, a need to make something meaningful, and a refusal to let work become hollow just because it sounds impressive.
So this is not meant to read like a traditional portfolio. It is closer to a self-portrait. Some of my life happens in models, agent systems, and research. A lot of it happens in long drives, honest conversations, family history, and the feeling that whatever I build should still feel human when it is done.
Still the same kid from Chennai, just building stranger things now.
I did not want to hide the personal side behind polished language. The products matter, but the person behind them matters too.
Chennai came first. Family came first. The ambition showed up later.
That part matters because it is still the foundation. My parents taught me to show up fully, not halfway. My family never had a shortcut culture. You stayed, you figured it out, you made it count. I carry that into everything, even the way I think about a problem at 2 a.m.
I am someone who will drive four hours just for a conversation that feels real, who still calls home more than most people my age, and who genuinely believes the best ideas come from paying attention to the people around you. I happen to build AI systems for a living, but the instinct underneath is the same one I learned at a dinner table in Chennai: care about what lasts.
Roads, water, long conversations, and my first motorcycle. That is where a lot of my clarity lives.
The technical side is real too. I have built ML models from scratch, fine-tuned multiple models, orchestrated agentic workflows that move across tools and decisions, and built infrastructure like a version-aware indexing layer for SDK docs and a temporal, knowledge-graph-based memory layer for AI agents.
I like hard technical problems, but I like them most when they are attached to something real and useful. That is usually the difference between work I respect and work I forget.
The through-line is simple: memory, initiative, systems that improve with use, and AI that feels less like a trick and more like serious infrastructure.
HiringBae is the company I keep returning to because it feels closest to what I believe. AI teammates should not just answer questions. They should remember what matters, communicate clearly, and carry meaningful work forward without needing to be babysat at every step.
BaeMemory came out of a frustration I kept seeing everywhere: agents could sound smart for a moment and then immediately lose the thread. I wanted memory to feel less performative and much more durable. That work lives most directly at baememory.com.
I keep getting pulled toward systems that learn from consequences, not just instructions. That difference matters to me because it is usually the line between a clever demo and something that actually grows sharper through use.
My research work through ASU comes from the same instinct, just in a more academic setting: use machine learning to speed up structural simulation problems that normally take far too long, while still staying close enough to the fundamentals to understand why the result works.
If you are building serious AI and care about memory, initiative, and trust, I would be glad to talk.
Say hello →