MS Statistics & Data Science
LMU SIST Scholarship Recipient
Grade: 1.8 ( 1 Highest )
Krrish Agarwalla
MS Statistics & Data Science Student
Munich, Germany
Hello 👋 I'm a Statistics & Data Science graduate student at Ludwig Maximilian University of Munich and SIST Scholarship recipient. I'm passionate about deep learning, language models, and building impactful technology. I love building products that solve real-world problems and I'm an open-source enthusiast. Previously founded Kaza Street Tech - an order management platform for street food vendors in India, which was recognized in Shark Tank Campus Special Round and selected as a Top 17 Student Led Startup from India.
Making sure that AI can on some feature phones
We are trying a new approach of distilling during pretraining for all 22 indic languages. We were able to compress models upto 8x and furthering this upto 20x in the future
Building language models that actually understand India's linguistic diversity.
Developing a Mixture-of-Experts architecture for morphologically rich, under-resourced Indic languages (Hindi, Marathi, Bengali, Tamil) with custom tokenization extensions and sparse MoE design for language-specific routing.
Digitizing 2000 years of legal history with deep learning.
Working with Dr. Zachary Chitwood and MLCU to extract Roman legal texts from medieval manuscripts written in Byzantine Greek using custom OCR pipelines and transfer learning for historical document processing.
Automating Germany's rental market with AI.
Building end-to-end ML infrastructure with David Bellem, Christian Deger, and Johannes Ditterich to automate property management—from document processing to tenant verification. Building production-grade LLM systems with full lifecycle from data collection to deployment. Avento
Making research reproducible, one paper at a time.
Automated tool for analyzing research paper reproducibility using conference guidelines (NeurIPS, ICML). Working with Dr. Anne-Laure & Dr. Moritz Herrmann on open science initiatives.
I regularly share my experiences, learnings, and insights on machine learning, language models, MLOps, and technology through my newsletter. Topics include deep dives into research projects, technical tutorials, and reflections on building systems that work in production.
Read My Newsletter