About

Hi! I am Manthan Gupta, an AI Research Engineer based in Bangalore, India. I spend my days diving deep into the systems that power modern AI, from how models run on GPUs to how agents orchestrate complex workflows.

What I’m Working On

Right now, I am knee deep in inference engineering, writing GPU kernels, and exploring distributed training. My focus in 2026 is on AI agents, distributed systems, inference engineering, LLMs, and GPUs, basically, everything that makes AI systems fast, scalable, and actually useful.

I thrive on solving hard problems and that feeling of being completely uncomfortable in a field I know nothing about. That’s where the best learning happens. My approach is simple: build to understand, write to clarify. If I can’t explain it clearly, I probably don’t understand it well enough yet.

Projects I Have Built

I have built a few things that might interest you:

• Water — A multi-agent orchestration framework that makes building agent workflows actually enjoyable. No more wrestling with verbose boilerplate.

• BYOM — Think of it as a USB for your memory layer. You own 100% of your data, no compromises.

• CricLang — A programming language for cricket enthusiasts, because why not combine two things I love?

What I Write About

This blog is where I break down complex topics—memory systems in AI agents, replication strategies in distributed databases, code search algorithms, and whatever else catches my curiosity. I write to clarify things for myself, and hopefully for you too.

When I Am Not Coding

I play cricket and badminton, watch F1 religiously, hit the gym regularly, and read way too many tech books. Balance is important, even when you’re obsessed with your craft.

Let’s Connect

I am always open to interesting conversations, collaboration opportunities, or consultation work. Whether you’re building something cool, need help with AI systems, or just want to chat about inference engineering, reach out!

Drop me a line on Twitter or email me at guptaamanthan01[at]gmail[dot]com.