ADITHYANADITHYANADITHYAN
MADHUMADHUMADHU

Software Development Engineer
Scroll
Adithyan profile photo

Backend Engineer focused on shipping useful systems.

I build reliable APIs, practical data models, and backend workflows that teams can use every day. My focus is clean architecture, performance, and developer experience without unnecessary complexity.

New Delhi, India · available for remote/hybrid work
ProjectsView all

VinciUI lets you wire together prompts, image inputs, and model settings through a clean node editor. Built for fast experimentation, it supports Google’s Gemini models (Flash 2.5, Flash 2.0 image preview) with a polished, minimal UI and GPU-friendly workflows.

ViteTypeScriptTailwindCSSNode Editor

Readingroom is a social media platform tailored for book enthusiasts. Users can track their reading progress, write reviews, post thoughts, and engage with other readers in a structured, AI-enhanced environment. Featuring personalized feeds, AI-curated recommendations, and clean UI, the platform is a digital home for bibliophiles.

Next.jsDjangoTailwindCSSPostgreSQL

Vichaar is designed to help you validate your ideas. Contain a huge libary of patent data which you can use to validate your ideas.

Vite(react)TailwindCSSPostgreSQLFastAPI
ToolkitDetailed view
Python
Django
FastAPI
TypeScript
React/Next.js
PostgreSQL
Docker
AWS
ExperienceFull history

Software Development Intern

Jan 2026 — Present
Paramount Products Private Limited

Built an internal platform managing 10,000+ garment designs and buyer workflows; shipped role-aware APIs that reduced manual processing by ~25%.

Software Development Engineer Intern

Apr 2024 — Oct 2024
Newgen Software Technologies Limited

Contributed to performance tuning, stability improvements, and UI/UX enhancements. Designed and integrated a RAG-based internal policy assistant.

Campus Outreach – Intern

Jun 2022 — Jun 2023
Geeks for Geeks

Hosted technical workshops for 300+ participants and mentored 50+ students on DSA and CS fundamentals.

Open Source

Intel Auto‑Round

Repo

Contributed to FP8 inference tooling: refactors to improve maintainability and PyTorch module traversal compatibility.

  • Refactored FP8 dequantization/detection using a registry pattern (merged PR #1348).
  • Migrated module access to PyTorch get_submodule/set_submodule APIs (merged PR #1365).
  • Collaborated with Intel engineers to improve readability, scalability, and backward compatibility.
PythonPyTorchFP8ML InferenceRegistry Pattern
Shipping Rhythm
PhotographyView gallery
Photo
Photo
Photo
Photo
Photo
Photo

Let’s build something meaningful.

Always open to opportunities, collaborations, and interesting conversations.

Say Hello