At Intel Unnati Industrial Training, I developed an image-similarity retrieval and recommendation pipeline using ResNet50 feature extraction, triplet-network-based learning, and FAISS nearest-neighbor search. I also implemented a FastAPI backend to connect image retrieval with the application interface.
I built VoiceFusion, a multilingual audio-visual dubbing pipeline combining transcription, translation, voice synthesis, and lip synchronization. My projects also include a speech-driven claim-verification application and a document-verification workflow for scholarship applications.

