Machine Learning Intern @ AI India Innovation
01 October 2023
[Enhancing LLM and OCR Capabilities]
From October 2023 to January 2024, I worked as a Machine Learning Intern at AI India Innovation Center in Pune, India, where I focused on fine-tuning large language models (LLMs) and optimizing OCR systems.
During my tenure, I fine-tuned open-source LLMs to answer high school math questions by generating synthetic data tailored to client requirements, achieving a model accuracy of 70%.
To enhance performance, I implemented quantization techniques to optimize LLM efficiency, successfully reducing model size while maintaining accuracy. Additionally, I developed and deployed an LLM-based solution with frontend integration, resulting in a fully functional application for end-users.
Beyond LLMs, I engineered an advanced image augmentation pipeline incorporating techniques such as image in-painting and image-to-image translation. Furthermore, I designed and implemented an OCR system for passport information extraction, automating data parsing and significantly improving efficiency.
