Tharun Kumar

Senior at Princeton University
I'm a research engineer interested in training large models to reason about the world.
Check out my blog!

I'm an engineer and researcher driven by a passion for building impactful AI systems that bridge cutting-edge research and real-world applications. Previously, I worked on state-of-the-art AI agents for Fortune 500 companies as a founding engineer at Universal AGI, increased availability of distributed systems at AWS, and researched generative video models at Hooglee under mentorship from Sebastian Thrun.

Currently, I'm a senior at Princeton University, studying CS and math. I'm looking for full-time opportunities in AI research and engineering. Aside from working on fun projects and learning as much as I can about the world, I'm looking for intense and fast-paced environments where I can ship products and push boundaries towards AGI.

In my spare time, I'm usually drumming, playing volleyball, lifting weights, or making new things (ask me about how I hem my clothes).

EXPERIENCE

Sep 25 - present

Research Engineer @ Google DeepMind

Researching and implementing novel State Space Model architectures using Spectral Filtering for efficient long-context language modeling under Prof. Elad Hazan in the Princeton Google DeepMind Lab. Training 500M, 1B, and 7B parameter LLMs from scratch, benchmarking performance against transformer and SSM baselines, achieving state of the art performance for language modeling in O(n) time (instead of transformers' O(n^2) time).

This research focuses on making language models more efficient at processing very long sequences of text, which is crucial for applications like document analysis and extended conversations.

PythonPyTorchState Space ModelsLLMsMachine LearningResearchSpectral Filtering
Jun 25 - Aug 25

Software Engineering Intern @ Amazon

Increased availability of large 8000+ server distributed system, increasing host redundancy by 15% and eliminating 30% of tickets for on-call engineers.

PythonTypeScriptAWSReactNode.js
Jan 25 - Apr 25

Member of Technical Staff @ UniversalAGI

Built AI solutions for Fortune 500 companies and governments, developing a state-of-the-art Text-to-SQL LLM agent with high accuracy on Spider 1.0 benchmark. Enhanced 3D Gaussian Splatting with Stable Video Diffusion optimization and created an LLM agent for Apple that builds no-code applications with external service integrations.

This was during my spring semester, so I was flying back and forth between San Francisco, UAE, and Princeton (for midterms and finals). I still ended up doing well in all of my classes :)

PythonLLMKubernetesText-to-SQL3D Gaussian SplattingStable Video Diffusion
View Full Resume

I love meeting new people. Feel free to reach out if you have questions or just want to chat :)

tharuntk (at) princeton (dot) edu