Education

• Emory University

  2025
  Master of Science in Computer Science
  GPA: 3.8/4
  
• Devi Ahilya University

  2023
  Bachelor of Engineering in Computer Engineering
  CGPA: 8.68/10; Class Rank: 3/150; graduated summa cum laude

Contact

• : prakashadityaap@gmail.com
• : GitHub
• : LinkedIn
• : Instagram
• : Stack Overflow
• : CodeChef
• : Leetcode

Research

• Quantum Circuits in Action: Defending Neural Networks from Adversarial Attacks

  • Evaluated adversarial robustness on the MNIST dataset by applying diverse attack methods (FGSM, PGD, SINIFGSM, VNIFGSM) using the Torchattacks library on both classical CNNs and Quantum Neural Networks (QNNs).
  • Designed and simulated 8 QNN architectures (linear, star, random, and full entanglement etc.) with Cirq, demonstrating that quantum models can exhibit up to 6.25% improved resistance against complex adversarial perturbations compared to baseline CNNs.
  • Developed a noise-augmented training strategy (using Salt-and-Pepper and Gaussian noise) that enhanced CNN accuracy by 3-10% and QNN accuracy by 2-6%, significantly bolstering model resilience against adversarial attacks.
  • Note: On-going project with plans to publish in a top journal.

  

• Concordance in Machine Learning and Deep Learning for Breast Cancer Screening

  • Enhanced tumor detection accuracy to 77% with 91% AUC by fine-tuning ResNet50 and InceptionNetV2 models for breast cancer screening, leveraging Java and Python for model optimization.
  • Processed and integrated 5TB of image and clinical data by creating end-to-end pipelines, ensuring seamless system integration and efficient data handling using ETL processes.
  • Improved diagnostic prediction by implementing Naïve Bayes, SVMs, kNNs, Decision Trees, and Logistic Regression, enhancing system integration and leveraging SQL for data manipulation.

  Code

  

Skills

Projects

• Canvas AI Assistant

  • Improved accessibility of AI-driven tools for students by developing a Python tool that creates an MCP-compliant server integrating AI assistants with Canvas LMS, structuring course data into knowledge graphs and storing it in SQLite.
  • Enhanced student experience by developing a syllabus parser that integrates with the Canvas API, automatically updating course modules, assignment due dates, and exam schedules and reducing manual tracking by 100%.
  • Built an AI-powered lecture material retrieval system, enabling instant access to summarized content and lecture insights, improving revision speed by 2× for the average student and ensuring comprehensive study resources for students.

  

• Intelligent Movie Recommendation System

  • Architected and developed a full-stack Movie Recommender System using Remix, React.js, Flask and PostgreSQL with advanced caching strategies using Redis.js to achieve an 85% cache hit rate, which reduced redundant database queries by 40%.
  • Integrated third-party TMDB API to aggregate a comprehensive movie database and implemented a custom indexing strategy that cached the 200 most frequently accessed movies, reducing query response times by 80%.
  • Achieved a 95% user satisfaction rating by conducting iterative usability tests, gathering actionable feedback, and refining both the interface and functionality to create a more intuitive UI/UX and engaging features.

  Code

  

• Sorting Visualizer

  A simple JavaScript application that visualizes sorting algorithms. Featuring a clean, intuitive, and minimalist UI, the application is easy to navigate and use. Built for use by developers and educators alike to enhance their own (or students') understanding of these algorithms.
  Code | Website