Open source · 2024
GPTs From Scratch
An open source, PyTorch-based implementation inspired by GPT models such as Llama and Mistral. Built to explore how these models work and the practical steps involved in training and deployment.
AI/ML systems architect · Pune, India
I build AI systems that are ready for the real world.
I work across the lifecycle of modern AI—from model training and fine-tuning to reliable, low-latency production systems.
Building, learning, and sharing along the way.
A little about me
Engineering ideas into useful systems.
I’m a Lead AI/ML System Architect at A10 Networks, working across the AI lifecycle—from model post-training and fine-tuning to the systems that deploy and protect AI applications. I’ve contributed to building AI Firewall and AI Gateway capabilities, connecting model expertise with reliable production infrastructure.
I also build open source projects and write about the systems and techniques I’m learning. I’m curious about how thoughtful engineering can turn ambitious ideas into products people rely on.
Things I’ve built
Projects & Open Source
Notes from the field
Writing
MCP and A2A Explained Simply: Two Concepts Everyone in AI Seems to Talk About
The Simplest Mental Model for AI Agents & Ecosystem Around It
GPUs, CUDA, Drivers and PyTorch: Concept, Setup & Installation
Kubernetes: The Mental Model That Makes Everything Click
Distributed Training, Part II: Choosing the Right Strategy — DDP, FSDP, TP+PP, and Beyond
Distributed Training Demystified: A Beginner’s Guide to DDP & FSDP
Building Production-Ready LLM Inferencing Pipeline: A Step-by-Step Guide
Demystifying LLM Training: A Programmatic Guide to Building and Training LLMs
Have something in mind?
Let’s connect and
build something useful.
Get in touch
Based in Pune, India