Varun Vasudeva

Code

software engineer

&

data scientist

Full Stack

Extensive experience building full-stack mobile & web applications using an array of technologies

Machine Learning

Proficiency building and using SoTA machine learning models for a variety of applications

Cloud Computing

Demonstrated expertise in building and deploying applications on cloud platforms

TypeScript
Python
Java
React and React Native
NextJS
Astro
TailwindCSS
Swift
TypeScript
Python
Java
React and React Native
NextJS
Astro
TailwindCSS
Swift
TypeScript
Python
Java
React and React Native
NextJS
Astro
TailwindCSS
Swift
TypeScript
Python
Java
React and React Native
NextJS
Astro
TailwindCSS
Swift

Professional

I started out building automations and dashboards with the goal of getting better visibility into client environments. Since then, I've had the chance to contribute to and eventually be a lead on multiple large-scale data consolidation programs that involved combining cloud and on-premises data. Lately, I've been developing frameworks in Python to leverage LLMs to accelerate the pace of everyday business processes for my team. Most recently, I have been architecting and building a datacenter network management tool.

Personal

I have both open sourced projects under my own name and under my company, Momenta Lab. Persimmon, my latest project, is research preprint browser - an iOS app geared towards casual readers and researchers wanting to get closer to understanding the latest breakthroughs in research. I also maintain an extensive homelab, a subset of which inspired me to write llm-server-docs, a comprehensive guide to private LLM inference with over 600 stars on GitHub.

Featured Projects

Persimmon: Research Simplified

2025

A performant, intuitive, private, AI-enabled browser for academic preprints on arXiv and SocArXiv. One-click summarize articles using your own LLMs, save your favorite interests and authors, and get notified about releases. A product by Momenta Lab (momentalab.com).

Download on the App Store

llm-server-docs

2024

Highly-starred, end-to-end documentation to set up your own local & fully private LLM server on Debian. Equipped with chat, web search, RAG, model management, MCP servers, image generation, and TTS.

View on GitHub

Wordplay

2024

Command line-interface (CLI) tool that lets users use language models to generate text-based games. It can use both Ollama for complete local model support and OpenAI-compatible cloud endpoints to create unique, fun games.

View on GitHub

Archive

2022

Highly personalizable, AI-powered journaling app that allows users to record their thoughts, feelings, and memories. Archive uses natural language processing to analyze logs and provide insights on their sentiments. It also features full-text search over entries, allowing users to access their logs quickly and easily.

Download on the App Store

Predicting Electron Mass Using Neural Networks

2021

Neural network that takes tabular data, collected at CERN's Large Hadron Collider and made available on Kaggle, such as electron spin, position, charge, etc. of a pair of electrons as input and predicts the invariant mass of the electron pair.

View on GitHub