Duration: June 2025 - Present
Description: Engineered and deployed production-grade AI pipelines by working closely on advanced projects, including Retrieval-Augmented Generation (RAG) systems and interactive AI voice agents. Leveraged the LangChain framework and vector databases to develop scalable architectures for efficient data retrieval and context-aware responses.
Duration: July 2025 - January 2026
Description: Accomplished high-quality RLHF prompt engineering and SFT fine-tuning by optimizing metadata-driven prompts, evaluating responses on key metrics including truthfulness, conciseness, and safety, and developing Chain-of-Thought (CoT) solutions to enhance model robustness against GPT and DeepSeek outputs.
Duration: May 2024 - October 2024
Description: Accomplished port logistics optimization and cargo security enhancement, contributing to a 30% reduction in truck turnaround time and a 40% improvement in cargo theft detection accuracy. Conducted in-depth port operation analysis, identified inefficiencies, and developed an OpenCV-based system for automated truck number plate recognition, container ID detection, damage assessment, and real-time security alerts for unauthorized cargo movement in warehouses. Also developed OCR technology for automatic extraction of container and truck number plates, integrated with IoT sensors for precise time logging and performance analysis, and worked with Generative AI and Large Language Models (LLMs) to support advanced AI applications.
Duration: May 2024 - June 2024
Description: Developed an interactive chatbot in Python using tkinter, integrating API calls for weather updates, news retrieval, and language translation. Additional features include joke generation, date retrieval, and mathematical operations. In parallel, implemented a context-aware Q&A admission chatbot using Flask, incorporating memory management for personalized responses based on user interactions, thereby enhancing user engagement. Lastly, the “Disease Prediction System using Machine Learning” project focuses on developing an intelligent system that predicts disease likelihood from health data using machine learning, aiming for early detection and proactive healthcare management.
Duration: March 2024 - May 2024
Description: During my internship at Null Class Pvt Ltd, I specialized in advancing emotion detection algorithms through a series of focused tasks. I compared time-series models with static image-based approaches, developed dynamic model adaptation for personalized predictions, and integrated cultural context to enhance cross-cultural recognition accuracy. Additionally, I created real-time emotion detection systems for streaming video data and trained multi-context models.