Responsibilities include:

• Led a cross-functional team to build an MVP chatbot for text2SQL applications using Langchain and Langraph. Improved system accuracy by conducting root cause analysis and redesigning schema by 27%, significantly enhancing SQL query generation and overall performance.

Responsibilities include:

• Launched ’Ditto,’ a project aimed at detecting similar software bug defects through semantic text similarity techniques. Enhanced the pipeline with a RAG (Retrieval-Augmented Generation) based system to optimize efficiency using Qdrant.
• Spearheaded the creation of an Interior Monitoring System for enhancing road safety, passenger comfort, and driver focus. Implemented advanced features such as Seat Belt Detection and Drowsiness Detection using Evidential Deep Learning. Leveraged Explainable AI algorithm GradCam++ to minimize False Positives and True Negatives, leading to a significant reduction in errors.
• Developed a smart system leveraging NLP and the MPNet sentence encoder model to associate bugs with test cases, achieving a 75-fold increase in efficiency and 95% test coverage.

Responsibilities include:

• Engineered an automated speech dubbing solution incorporating ASR, Speaker Diarization, and TTS capabilities, leading to a more accurate and natural-sounding output.
• Customized a TTS synthesis model using the Indic TTS dataset, effectively capturing regional accents and dialects.

Responsibilities include:

• Crafted an offline wake-up word detection system utilizing GRU Networks for real-time responsiveness, significantly reducing latency by 13%.
• Formulated a Siamese Deep Neural Network for facial recognition, achieving a 99.8% accuracy rate, contributing to enhanced security measures.