Adarsh Dubey

I'm a web developer and machine learning engineer who loves building in public. Currently contributing to Google DeepMind through GSoC '25, and lately, I've been tinkering with LLMs. I speak at local tech meetups, write technical blogs, and when I'm not lost in code, I'm probably enjoying misal pav.

Career highlights

Google DeepMind

Google DeepMind

GSoC '25 contributor

May 2025 — Aug 2025

WeMakeDevs

WeMakeDevs

Web developer & core team member

2023 — Present

Projects

Facet: Gemma Finetuner

An end-to-end platform that lets users preprocess datasets, fine-tune Gemma models, run inference, and export them in GGUF format, making model training and deployment simple and accessible.

Python

Python

FastAPI

FastAPI

Hugging Face

Hugging Face

Docker

Docker

GCP

GCP

Next.js

Next.js

Anon: Anonymous social media forum

Anon Forum is a fully anonymous social platform for college students that enables posting, deleting, upvoting via an optimistic UI, and ensures anonymity through college verification with no personal data storage.

Next.js

Next.js

FastAPI

FastAPI

AWS

AWS

PostgreSQL

PostgreSQL

TypeScript

TypeScript

Bun

Bun

Inclinet: Neural network library

Built a neural network library from scratch in Python with PyTorch-like APIs, supporting multiple activation functions, losses, and optimizers, designed using OOP principles for clarity, flexibility, and maintainability.

Python

Python

Incligrad: Autograd engine in C++

Implemented a PyTorch-like autograd library in C++ from scratch, including full backpropagation logic, enabling gradient-based learning without relying on existing frameworks.

C++

C++

Recent blogs

all blogs

Technologies

Python

Python

PyTorch

PyTorch

Keras

Keras

Hugging Face

Hugging Face

C++

C++

React.js

React.js

Next.js

Next.js

TypeScript

TypeScript

Bun

Bun

FastAPI

FastAPI

PostgreSQL

PostgreSQL

Git

Git

AWS

AWS

GCP

GCP

Docker

Docker

Gemini

Gemini

Claude

Claude

Stay connected

You can put in your email below to get notified when I publish new blogs. No spam, promise :D