Kashish Khatri

Khoury College of Computer Sciences | Expected graduation: May 2027

Master of Science in Computer Science | GPA: 3.33/4.0
Related courses: Programming Design Paradigm, Algorithms, Databases, Machine Learning

Bachelor of Computer Applications (BCA) | GPA: 8.98/10.0 (Top 3%)
Related courses: Data Structures, Java, Python, Algorithms, Operating Systems, Databases, Web Development

Software Developer Intern

• Achieved 40% reduction in content search time by developing Book-E-Pedia using React.js, Django REST, and PostgreSQL, improving platform usability.
• Boosted user engagement by 25% by integrating four media formats (physical books, audiobooks, e-books, and video books) into a single responsive interface.
• Delivered 50% faster report preparation through a Customer Insights Dashboard built with React.js, Django REST, and PostgreSQL to visualize orders and track customer trends.
• Enabled 100% digital analytics distribution by implementing CSV and PDF export functionality using PapaParse and jsPDF, automating report generation.

• Integrated 6 Azure AI services (Computer Vision, OpenAI, Speech, Maps, Communication Services, Translator) into a full-stack healthcare platform built with Django REST Framework and React.js.
• Reduced manual prescription entry by 80% through Azure Computer Vision OCR and OpenAI-powered drug interaction analysis, improving medication safety for patients and pharmacies.
• Supported 5 distinct user roles (patients, doctors, pharmacies, laboratories, admin) with automated medication reminders, lab report processing, and real-time translation across 10+ languages.

• Reduced video creation time from hours to under 90 seconds by building a two-stage AI pipeline using Stability AI's SDXL and Stable Video Diffusion models via Replicate API.
• Achieved 95% API reliability by implementing intelligent rate limiting with exponential backoff retry logic, handling 100+ daily generation requests.
• Designed responsive glassmorphism UI supporting 3 motion intensity levels, real-time progress tracking, and generation history for improved user experience.

• Conducted research on skin disease prediction using Convolutional Neural Networks (CNN) and Softmax, achieving an accuracy of 87.6% through extensive training and validation on a custom-built dataset.
• Collected and cleaned over 5,000 images via web scraping to create a high-quality dataset, significantly improving model performance.