About
As my freshman year was getting over, my interests shifted from cs fundamentals to what really made those binary 1s and 0s do what they do. With a perfect summer oppurtunity to build upon my skills through projects, I did not even realize how fast I went from tutorial videos to deep research papers. Voila! now I have researched at multiple labs, interned at a few places, and built some cool projects.
Work Experience
- W
WüNLP, University of Würzburg
Research Intern • Onsite
Filtered reasoning-intensive samples from 10+ public datasets using LLM-as-a-Judge techniques, and deployed a scalable vLLM-based synthetic data generation pipeline on SLURM clusters to produce over 2 million training samples. Evaluated performance across 15+ SOTA LLMs for reasoning tasks. - I
IN3TECH Computer Systems - FZE, Dubai
Software Engineer Intern • Remote
Built a Meta Glasses-powered face recognition system integrated with a Flask web app and Google Search API for enriched personal data retrieval using LLMs like LLAMA and Gemini. Also developed a paperless parking solution with 82% accuracy using TensorFlow, OpenCV, and MySQL-backed Flask APIs. - T
TavLab, IIIT Delhi
Undergraduate Researcher • Onsite
Constructed a temporally-aligned clinical dataset from the MIMIC database, increasing lab test coverage from 4 to 20. Enriched predictions using SciSpaCy embeddings from unstructured notes and trained LightGBM/XGBoost models achieving over 80% accuracy within 20% deviation. - M
MIDAS Lab, IIIT Delhi
Undergraduate Researcher • Onsite
Worked on automated code generation, creating a 470K-sample dataset and an extensive evaluation pipeline, and co-authoring a survey paper. Also worked on a suspect retrieval system for Delhi Police training FIRe model on CelebA and multimodal LLaVA model combining images and descriptions to improve identification accuracy.
Here is my work
Though just a handful, these projects cover everything from customer care solutions to personalized health recommendations.
RecommendCare
Engineered a personalized health platform using Apple Watch data, hybrid prediction models, and RAG-based LLMs for real-time vitals monitoring and tailored health recommendations.
Drop Me a Line
Interested in connecting? Shoot me a message on LinkedIn or email me directly, and I’ll be right at ya!