Prachi Chavhan

About

Computer Engineering student at VJTI Mumbai with hands-on experience building full-stack web applications and machine learning projects. Interested in software development, data-driven systems, and applied AI.

Education

• HSC (PCM)
  New English Junior College
  PCM · 2021 – 2023
• B.Tech in Computer Engineering
  Veermata Jijabai Technological Institute (VJTI)
  Computer Engineering · Aug 2023

Skills

Tools / apps / platforms

Projects

• Multimodal Fake News Detection ↗
  Python, TensorFlow/Keras, OpenCV, Transfer Learning

  Developed the deepfake detection module of a multimodal fake news detection system using ResNet50 with transfer learning. Preprocessed and augmented the Human Faces dataset using resizing, normalization, rotations, flips, and zoom to improve model robustness. Trained and evaluated the deepfake detection model using binary cross-entropy loss and Adam optimizer, achieving 95.5% accuracy. Collaborated with teammates to integrate the vision module into a larger multimodal pipeline combining OCR, CLIP-based text-image consistency, and image forgery detection.

• Driver Drowsiness Detection ↗
  Python, PyTorch, OpenCV, Scikit-learn, ResNet-18

  Developed the ensemble inference pipeline by integrating a ResNet-18-based CNN with a classical machine learning model using facial landmark features. Combined predictions from geometric features (EAR, MAR, Eye Circularity, MOE) and deep learning models to improve robustness, especially under eye occlusion such as sunglasses. Evaluated model performance on the YawDD dataset and compared individual model performance against the ensemble approach. Implemented a real-time webcam-based inference workflow using OpenCV for low-latency driver alertness monitoring.

• College Discovery Platform ↗
  Next.js, TypeScript, Node.js, Express.js, PostgreSQL, Prisma, JWT

  Independently developed a full-stack college discovery platform enabling students to explore, compare, and shortlist over 1,200 colleges using a real-world dataset. Designed and implemented RESTful APIs for search, filtering, comparison, authentication, and saved colleges using Express.js, Prisma ORM, and PostgreSQL. Built responsive interfaces with Next.js, TypeScript, and Tailwind CSS, including dynamic college pages and side-by-side comparison of up to three institutions. Developed a CSV-based data import pipeline with validation and normalization for reliable PostgreSQL data storage and deployed the application using Vercel and Render.

Courses & certifications

• Getting Started with Deep Learning ↗ · NVIDIA

⚽ Extracurricular activities

• Hackathon Team Lead

  Led a four-member team during a hackathon, coordinating development, task allocation, and final project delivery within strict deadlines.

• Mentor, Inheritance CoC, VJTI

  Mentored junior students on Machine Learning projects, providing guidance on implementation, debugging, experimentation, and best practices.