MS CS at Rutgers University (2024) with specialization in AI || Ex Software Engineer at Soroco

Shashwath Santhosh  • Peerlist

• Architected a computer vision solution to identify under-extrusion errors in 3D printing processes, integrating a fine-tuned ResNet-50 model with LDS edge detection to enhance spatial feature extraction. Tailored for high-throughput analysis and processed through rigorous validation and testing protocols to refine accuracy.
• Achieved a precision of 75%, ranking 7th on a competitive leaderboard, showcasing the model's capability in real-time anomaly detection and potential for deployment in industrial quality control systems.

3D Printer Error Detection

•	Engineered a specialized chatbot for diagnosing skin conditions, utilizing Retrieval Augmented Generation (RAG) with a Llama-index backend on AWS. This system featured a Streamlit interface, enabling real-time user interactions and improving response accuracy by 20%.
•	Designed to support multimodal inputs, allowing inquiries using both text and images, thus expanding the assistance scope and enhancing diagnostic accuracy by 25%.

Skin Condition Chatbot with RAG

•	Contributed in a group of 12, to create a real-time order tracking application in a Single Page Application, using Node.js with MySQL, hosted on AWS, handling data for over 500 restaurants and 2,000 daily orders.
•	Incorporated third-party APIs for payment processing and location services, such as Stripe, Google Maps and bcrypt, and crafted comprehensive test cases to ensure functionality and reliability, achieving 98% test coverage.