At NIT Warangal, I extended the SIGIR 2025 DGMRec model with a module to improve recommendation quality when modality data is missing, especially for new items. I tested it across Amazon Clothing, Sports, and Baby datasets, with the best Sports results reaching Recall@20 0.1020 and NDCG@20 0.0455.
I also authored a research manuscript with a 20-baseline comparison and ablation study. My work included 30 tuning experiments to optimize the model’s key hyperparameters.
I built a generative AI chatbot with persistent SQLite memory, using conversation history for multi-turn context and a Streamlit interface. I also developed a voice assistant pipeline for speech recognition, command processing, and text-to-speech.
For a deployed movie recommendation project, I used TF-IDF, Bag of Words, and cosine similarity to generate content-based recommendations. I also developed an internship and placement analytics system that combined Python data processing with a Power BI dashboard for KPI tracking, placement trends, and skill demand analysis.

