Prakhar Mishra

Developed a Natural Language Generation (NLG) model to generate concise summaries from any topic using NLP and sequence-to-sequence architectures. Optimized model performance and collaborated on real-world deployment for improved accuracy and readability.

Pathfinder for UPES (Minor Project)

Developed a campus navigation tool using C++, Node.js, and HTML/CSS that reduced navigation time by 30% and integrated real-time event/construction updates.

Movie Recommender System

Built a content-based recommendation engine with Python, Scikit-learn, and Flask, achieving 85% accuracy in personalized movie suggestions.

Vision – OCR Powered(Minor Project)

Designed an OCR system using Python, OpenCV, EasyOCR, and Tesseract to extract text from images/PDFs and classify AI vs human-generated text with BART-large MNLI.