Projects
Artificial Intelligence (AI) Image Pipeline
Artificial Intelligence (AI) Consultant
Led the creation and deployment of an AI image processing pipeline for packaged food products, achieving a 95% accuracy in detecting ingredients, allergens, logos, and more. The system integrated advanced technologies like GPT-4 and Google Cloud Vision OCR, enhancing data extraction accuracy and text recognition. The project also involved the successful identification of logos across thousands of brands, comprehensive DevOps management, and the establishment of robust database structures. Through these innovations, manual interventions were entirely eliminated, and system uptime was maximized.
Advanced Deep Learning Chatbot
Associate Data Scientist
Specializing in NLP, machine learning, and web development, the project centered on the development of a deep learning chatbot that enhanced user interactions, reducing query resolution time by 35%. An analytics dashboard was introduced, offering real-time insights and data visualizations. The chatbot was seamlessly integrated into collaboration tools, boosting response times and team efficiency. Additionally, an administrative console was designed, optimizing task execution and minimizing manual interventions.
Racetrack Reinforcement Learning Simulation
Research Scholar
Developed a gridworld simulation of a race car navigating a track, with specific start, finish, valid, and boundary positions. Implemented an on-policy Monte Carlo method to determine the optimal racing policy, achieving 95% successful track completions. Utilized an off-policy Monte Carlo approach to further optimize the racing strategy. Analyzed and compared the convergence properties of both methods, showcasing trajectories and state value examples for deeper insights.