Chemical Engineering BTech & Data Science MTech
Indian Institute of Technology, Madras (CGPA: 9.04)

I am a passionate technologist with a strong affinity for problem-solving and leading teams toward innovative solutions. As a free thinker, I thrive on exploring out-of-the-box ideas and cutting-edge technologies like Large Language Models (LLMs) and Multi-Modal LLMs. My journey at IIT-Madras, coupled with hands-on experience in AI, data science, and software development, has equipped me with the skills and mindset to tackle complex challenges.

• Programming Languages: Python, JavaScript, MATLAB, SQL, C, C++
• Frameworks & Libraries: PyTorch, TensorFlow, ReactJS, Flask, FastAPI, HuggingFace, LangChain, PEFT, OpenCV, FastAI, Haystack-NLP, Scikit-learn, XgBoost
• Tools & Software: Simulink, MS-Office, Docker, Apache Beam, MLflow, Airflow, Prometheus

• Dual Degree Project | AI4Bharat Lab: Working on Indic LipSync generation using Wave2Lip and Super Resolution, and developing datasets and models for deep fake detection and multi-modal classification.
• Pulsar Detection (MLOps Project): Built an MLOps pipeline for pulsar detection, integrating FastAPI, Docker, and Prometheus.
• Identifying Chemical Kinetics: Applied Physics-Informed Neural Networks and attention mechanisms to chemical kinetics problems.

• Markdown Code Runner: Published on PyPI: A tool for executing code blocks within Markdown files, supporting languages like C, C++, and Python.
• Exploring Attentions: Implemented various attention mechanisms and created a comprehensive tutorial for understanding their operations.

• Inter IIT Tech-Meet-12.0 | Adobe (Rank 3 - Bronze Medal): Developed multimedia retrieval systems using language embeddings and fine-tuned LLMs.
• American Express Campus Super Bowl (2nd Runner-Up): Implemented a 4-level deep stack ensemble for default prediction, achieving top accuracy.
• AstraZeneca AI Hackathon | Winner: Trained a CNN model on a large-scale dataset, achieving high accuracy in document segmentation.

• RE-GAINS & EnCHANT (arxiv): Intelligent tool manipulation systems for enhanced query responses using API calls and tool chaining.
• Provisional Patent: Data Generation Techniques for Tool-based LLM Applications - Innovative pipelines for generating realistic tool manipulation data.

• Deputy Contingent Lead | Inter IIT Tech-Meet-12.0: Guided teams to success, securing multiple medals in AI and product development challenges.
• AI Hackathon Lead | AI Club, CFI: Pioneered the AI Club's establishment, fostering a tech community focused on problem-solving and innovation.

Feel free to explore my repositories and projects. I am always open to collaboration and discussions on technology and innovation!