Aasheer Magendran

Designed and implemented a web-based class path generator to assist community college students in planning their academic and transfer journeys.

• Designed system architecture integrating APIs like ASSIST and Degree Works for personalized academic planning.Built modular tools to accommodate university-specific pathways, class overlaps, and prerequisites.
• Engineered algorithms to optimize course selection based on user-defined transfer goals. Automated data scraping and API interactions using Selenium, Requests, and BeautifulSoup.
• Processed and managed data using JSON structures to handle class dependencies and requirements. Automated scheduling of academic events for streamlined planning. Delivered a comprehensive project demonstration, emphasizing real-world impact and usability.

Garbage Segregation System with YOLOv5

• Developed a high-accuracy waste segregation system using OpenCV, PyTorch, and YOLOv5. Built a labeled dataset with Grad-CAM, SAM, and active learning and automated real-time image classification and preprocessing.

Selected participant (1 of 100 chosen from 1,300+ applicants)

• Collaborated in a week-long program focused on AWS cloud services and AI/ML applications, gaining hands-on experience with Bedrock, S3, DynamoDB, Lambda, IAM, and SageMaker.
• Developed a vendor procurement dashboard in a hackathon-style challenge, enabling universities to manage, track, and analyze purchases .
• Designed and integrated AI agents powered by large language models (LLMs), applying prompt engineering techniques to optimize performance.
• Contributed to full-stack development efforts, spanning backend workflows, AI integration, and deployment.