An advanced conversational agent designed to guide students through complex math subjects. Implements sophisticated RAG pipeline with dynamic tool-calling, Wolfram Alpha integration, and Manim-generated explanatory videos with TTS.
Automated media intelligence pipeline for processing short-form video content. Video ingestion, frame/audio extraction, semantic classification using OpenAI CLIP, and speech transcription using NVIDIA Parakeet-TDT.
Robust optical character recognition system optimized for document intelligence. Achieved significant improvements in structured data extraction through model selection and hyperparameter tuning.