Taaha Khan
Hi, I'm Taaha!
I'm a Computer Science student at Purdue University with a strong interest in artificial intelligence and research-driven problem solving. My work spans areas such as large language models (LLMs), retrieval-augmented generation (RAG), compression, image processing, and game AI. I'm particularly interested about exploring tree search algorithms, information retrieval, and deep learning techniques to build efficient, intelligent systems.
You can contact me at taahajkhan [at] gmail [dot] com
website,
github,
linkedin,
kaggle,
google scholar
Projects
Enhanced LLM Reasoning with MCTS
-
Enhancing Mechanistic Interpretability and Multi-Step Logical Reasoning in Large Language Models
with Low Rank Adapted Self-Distillation from Monte Carlo Tree Search Chains of Thought.
[presentation]
CompaCT Image Compression
Research
Oregon Health and Science University, Casey Eye Institute, Dept. of Ophthalmology
- Dynamic traction force in trabecular meshwork cells: A 2D culture model for normal and glaucomatous states [publication]
- Comparative analysis of traction forces in normal and glaucomatous trabecular meshwork cells within a 3D, active fluid-structure interaction culture environment [publication]
- Calculating Traction Forces in Normal and Glaucomatous Trabecular Meshwork Cells in an Active 3D Fluid-Structure Interaction Culture Setting [publication]
NASA, UT Austin Center for Space Research, Cockrell School of Engineering
- Geostationary Orbit (GEO) Satellite Compliance Modeling with International Telecommunication Union (ITU) Regulation [publication]
Vital Signals Health
- Camera-based PPG Signal Extraction with Bidirectional LSTM Autoencoders for Low-Cost Vital Health Monitoring [website]
- Patent in preparation
Interests
AI, LLM, RAG, RL, DL, ROI, more acronyms...