Chirag Jagad

Chirag Jagad

Research Enthusiast | MS CS @ USC

Sitemap

A list of all the posts and pages found on the site. For you robots out there is an XML version available for digesting as well.

Pages

About me

Posts

experience

Data Scientist - Graduate Research Assistant

Keck Medicine of University of Southern California | May 2025 - Present

• Developed autoencoders and latent diffusion models using MONAI for high-quality brain MRI synthesis • Implemented DenseNet model in PyTorch for Alzheimer’s classification using brain images, achieving an AUC of 0.92 • Created regression models for genomic analysis, achieving 90% correlation in identifying genetic markers of disorders

STATimMED Research

Data Analyst Intern | May 2025 - Present

• Derived actionable insights from large healthcare data, ensuring HIPAA compliance and enhancing client decisions • Optimized SQL queries & built data pipelines to improve efficiency and reduce runtime, enabling faster data processing • Applied advanced statistical, data analysis, & data modeling techniques using Python, R, Excel & BigQuery to conduct comprehensive analysis, develop decision strategies, & create visualizations, driving informed business decisions

Dwarkadas J. Sanghvi College of Engineering (DJCSE)

Research Assistant | August 2021 - January 2023

Advisor: Prof. Neha Katre

• Developed a novel key-frame extraction algorithm for video surveillance system & optimized the run time by 97.5% • Employed a custom-trained YOLOv8 model with an accuracy of 95% to detect anomalies like fire and weapons • Authored a paper about video analysis and performed a thorough performance analysis for object detection models • Created a custom YOLO fire & weapon dataset specifically for detection from CCTV footage with low resolutions

University of California, San Diego (UCSD)

Research Intern | August 2021 - February 2023

Advisor: Prof. Pengtao Xie

• Collaborated and implemented an AutoML framework using PyTorch for decomposing object features into domain-specific & class-specific representations using neural architecture search approaches to solve the problem of domain adaptation • Optimized the framework and employed ResNet architectures with supervised & semi-supervised settings, & achieved a state-of-the-art accuracy of 87.5%

DotSyndicate

Software Engineer Intern | March 2021 - February 2022

• Designed an end-to-end admin portal using JavaScript & Node.js for Hustler Tanzania, an e-commerce company • Built an efficient data management system & triggers to handle the amount & velocity of data of company’s products and users • Provided real-time data analytics using Python for daily business analysis and presented strategic business decisions

Indian Institute of Technology, Delhi (IITD)

Research Intern | March 2022 - January 2023

Advisor: Prof. Tapan Gandhi

• Tasked on quantifying the distribution of magnetic susceptibility within brain MRI images using deep neural networks • Implemented and boosted the accuracy of the U-Net model built using TensorFlow to 91% by leveraging the generated data with 3D geometrical shapes for training, & improved the model by applying structural similarity index loss function

projects

FindIt – A solution for lost and found items

Full Stack Web Development

  • Constructed the backend using Node.js for an application that allows civilians to upload the data of lost items and search for their lost items.
  • Formed an admin portal using React and Node.js for organizations, that have lost and found departments, where admins can perform verification of the lost items and give them to the owners.
  • Included key features that include a real-time chat box, rightful owner verification, and data analytics

E-commerce Website

Full Stack Web Development

  • Developed a fully functional e-commerce website with end-to-end functionality and designed the database schema and managed the database
  • Built and designed algorithms for key features like Advance Searching, Cart management, PayPal payment integration, Wishlist, Order Tracking, and Reviews and Ratings

DJCSI Official Website

Full Stack Web Development

  • Built the official website of DJCSI (The official student chapter of Computer Society of India at my college) using React.js

Federated Machine Learning based Bank-Customer Churn Analysis

Machine Learning

  • Built a robust Multilayer Perceptron model using the Federated Machine Learning approach providing an accuracy of 96% for Bank Churn Prediction and solved the problem of the imbalanced dataset used by upsampling it
  • Presented a deep analysis of the advantages of Federated Machine Learning concerning the accuracy, transmission time, and computation cost

Car Damage Detection

Machine Learning, Computer Vision

  • Developed an intelligent system using TensorFlow to predict the repairing cost of the damage by locating and recognizing the severity of the damage.
  • Employed transfer learning and fine-tuned the VGG16 model to achieve an accuracy of 86%.

Smart Healthcare System

Machine Learning, Computer Vision

  • Developed an end-to-end website and an application to make technology more useful in solving problematic healthcare challenges by diagnosing various symptoms, analyzing data, and then predicting the diseases
  • Built models to predict Myopia, Color Blindness, PCOD, and to generate heart rate using head motions in video and predict the probability of Coronary Heart Disease

publications

Intelligent Car Damage Detection System

Publication: Springer Lecture Notes in Electrical Engineering, 2023

Current status: Presented at 2nd International Conference Women Researchers in Electronics and Computing

talks