Holocron is an educational platform that provides personalized learning roadmaps tailored to individual goals. It simplifies complex pathways like Python, Cybersecurity, and Machine Learning, guiding users step-by-step to mastery through adaptive recommendations and progress tracking.
It's a project that's designed to enable GPT creators to showcase their custom GPT's and get recognised through my GPT's store.
Semantic Book Recommendation is an AI-powered system that suggests books based on their semantic meaning using Large Language Models (LLMs) and vector embeddings. It transforms book descriptions into mathematical representations for efficient similarity search via ChromaDB. The system incorporates NLP techniques like zero-shot classification and sentiment analysis to enhance recommendations, providing context-aware book suggestions beyond simple keyword matching
A high-performance CUDA implementation of matrix multiplication that achieves near-cuBLAS performance through various optimization techniques. The project implements tiled matrix multiplication with shared memory, thread coarsening (2 elements per thread), and proper memory coalescing. The implementation handles large matrices (4096x4096) efficiently and demonstrates advanced CUDA optimization principles including proper boundary checking and shared memory padding to avoid bank conflicts.