SHREYAS DEB

Ambitious Computer Science undergraduate from Sikkim Manipal Institute of Technology with a solid foundation in programming languages including C, C++, Java, and Python. Experienced in Data Structures and Algorithms, with a strong interest in software development. Pursuing summer internships in AI or backend development to apply academic knowledge in industry settings.

Abdominal Organ Segmentation using nnU-Net
Tech: Python, PyTorch, MONAI, Deep Learning, Medical Imaging

• Led a team in implementing and training the nnU-Net model for automated segmentation of abdominal organs
• Preprocessed and augmented 3D medical imaging datasets to optimize training accuracy
• Contributed to project-based learning by mentoring peers and documenting the pipeline for future use
• Integrated into a broader research goal of enhancing diagnostic support using AI

Academic Tracker Application
Tech: Python, PyQt5, MySQL, OOP, SQL

• Designed and developed a cross-platform desktop application for students to track attendance, grades, and GPA.
• Built a user-friendly interface using PyQt5 with real-time dynamic dashboards and secure login authentication.
• Implemented efficient MySQL backend for data persistence and management.
• Helped over 20+ peers manage academic records during testing phase, receiving positive usability feedback.

DBMS.AI (Ongoing)
Tech: AI, Natural Language Processing (NLP), SQL, Python

• Currently designing an AI tool that converts natural language queries into accurate executable SQL statements.
• Built for non-technical users to query databases using everyday language with instant output rendering.
• Leveraging NLP and rule-based parsing for accurate intent detection and schema mapping.
• Incorporates SQL injection prevention and error correction for robust query handling.
• Goal: democratize data querying and enhance database accessibility.

Stock Market Analyser (Ongoing)
Tech: AI, Meta-Learning, Python, Financial Analytics

• Developing a meta-learning model to aggregate predictions from multiple AI algorithms (Random Forest, LSTM, XGBoost).
• Aims to enhance the accuracy of stock market trend predictions by combining model outputs
• Employing feature engineering on historical stock data to improve forecasting reliability
• Expected to outperform individual model baselines by 10–15% based on simulation data
• Project focuses on financial data science and real-world AI deployment strategy

Team Leader, Smart India Hackathon 2024
Led a team of five in developing a Grievance Management System

• Spearheaded the design and implementation of a user-centric grievance management system to streamline complaint handling
• Coordinated team roles, project planning, and presentation to judges during the college-level hackathon

Technical Secretary, Access (All Core Computer Engineering Student Society)

• Facilitated technical events, workshops, and collaborative student projects
• Acted as a liaison between faculty and students to promote a culture of innovation and continuous learning

Project Based Learning leader, abdominal organ segmentation initiative

• Guided a hands-on deep learning project using nnU-Net to segment abdominal organs from medical scans.
• Led dataset preprocessing, model training, and performance evaluation
• Empowered peers through structured documentation, walkthroughs, and result discussions, aligning with modern experiential learning models