Prashanna Raj Pandit

Prashanna Raj Pandit

Graduate Research Assistant

Computer Science Student

Graduate Research Assistant at Southern Illinois University Edwardsville specializing in Data Analytics, ML, Computer Vision, and Deep learning. Passionate about leveraging AI for automation, pattern recognition, and solving real-world problems.

About Me

I am a Computer Science graduate student at Southern Illinois University Edwardsville (SIUE), specializing in Statistics, Machine Learning, Deep Learning, and Computer Vision. I am originally from Nepal, where I completed my Bachelor’s degree in Electronics, Communication, and Information Engineering from Tribhuvan University, Pulchowk Campus—one of Nepal’s leading engineering institutes.

Currently, I am working as a Graduate Research Assistant, focusing on developing deep learning models for multimodal time-series data, particularly in gait biomechanics and noisy real-world datasets. My work involves end-to-end data processing, model design, and high-performance computing on GPU clusters.

I have hands-on experience with TensorFlow, Keras, PyTorch, and OpenCV, and a strong foundation in predictive modeling, data preprocessing, and statistical analysis. My goal is to bridge theory and real-world applications through intelligent, data-driven systems.

Technical Skills

Programming Languages

Python C C++ R SQL

Frameworks & Libraries

TensorFlow PyTorch Scikit-Learn Keras OpenCV YOLO Flask Pandas NumPy Matplotlib Seaborn Selenium

Tools & Software

Git/GitHub Docker Linux Tableau Curser PostgreSQL Google Colab Mlflow

Concepts

Machine Learning Statistical Modeling Data Preprocessing Neural Networks/DL Time Series ML Pipeline Model Deployment Cloud Services

Experience

Research Assistant

Jan 2025 – Present

Southern Illinois University Edwardsville

Edwardsville, IL, USA

Research in gait analysis and AI-based motion understanding using deep learning and signal processing.

  • Extracted 2D keypoints from 106 video datasets using OpenPose and processed JSON data with Pandas and NumPy to compute gait kinematic features such as stride length, cadence, and joint angles, demonstrating robust data preprocessing methods for scalable AI model development.
  • Engineered a multimodal dataset for deep neural networks by preprocessing gait time-series data via cubic spline interpolation, Savitzky-Golay smoothing, gait phase detection, and turn removal, aligning with best practices in feature engineering and data preparation.
  • Developed a logistic regression baseline achieving 72% accuracy and a deep learning model (80K parameters) achieving 88% accuracy on gait classification, showcasing experience in building and optimizing machine learning solutions.

Software Engineer

Jan 2023 – Jul 2024

Vrit Technologies

Kathmandu, Nepal

Full-stack web development and backend optimization for scalable AI-ready systems.

  • Designed and deployed the FootBalance-Nepal website with a Flask backend and Sheety API integration, improving SEO score from 71 to 88 and performance score from 81 to 95, demonstrating expertise in scalable web deployment.
  • Collaborated on full-stack development and optimized backend workflows for scalability and maintainability, reflecting an agile approach and readiness to integrate AI/ML components within larger systems.
  • Utilized Git for version control to ensure code consistency and facilitate team collaboration, supporting robust software development processes similar to MLOps workflows.

Projects

iGAIT: Gait Analysis and AI-Based Motion Understanding
Deep Learning / Computer Vision

iGAIT: Gait Analysis and AI-Based Motion Understanding

AI-based gait analysis using deep learning and signal processing for motion understanding.

Python OpenPose NumPy
Music Generation using RNN (Let's Dance!)
Deep Learning / Sequence Modeling

Music Generation using RNN (Let's Dance!)

Character-level music generation using LSTM-based RNNs trained on symbolic music data.

Python PyTorch RNN
Debiasing Facial Detection Systems
AI Ethics / Computer Vision

Debiasing Facial Detection Systems

Implemented CNNs and DB-VAE for bias mitigation in facial detection systems.

Python TensorFlow VAE
Advanced AI Project: Reinforcement Learning
AI / Reinforcement Learning

Advanced AI Project: Reinforcement Learning

Implemented Q-learning and Value Iteration agents for policy optimization and exploration.

Python Reinforcement Learning Q-Learning
MedNLPify: Deep Learning for Biomedical Text Classification
NLP / Deep Learning

MedNLPify: Deep Learning for Biomedical Text Classification

NLP system for sentence classification in biomedical abstracts using hybrid embeddings.

Python TensorFlow Flask
Food Vision Project
Computer Vision

Food Vision Project

Food classification model using transfer learning on the Food101 dataset.

TensorFlow CNN Transfer Learning
ExFS2: User-Space Virtual File System
Systems Programming / File Systems

ExFS2: User-Space Virtual File System

Implemented an inode-based virtual file system supporting direct and indirect addressing.

C File Systems Memory Management
Traffic Sign Classification with Fine-Tuned VGG-16
Computer Vision

Traffic Sign Classification with Fine-Tuned VGG-16

Traffic sign classification using fine-tuned VGG-16 with TensorFlow.

TensorFlow VGG-16 Transfer Learning
User-Level Threading and Scheduling
Operating Systems

User-Level Threading and Scheduling

Implemented many-to-one user-level threading with Round Robin and Lottery schedulers.

C ucontext Operating Systems
Delivery Route Planning Optimization
Algorithm Design / Optimization

Delivery Route Planning Optimization

Optimized e-commerce delivery routes for Chicago using advanced algorithms.

Python NetworkX OpenStreetMap
Car and Sidelight Detection with YOLOv5
Computer Vision

Car and Sidelight Detection with YOLOv5

Real-time detection of cars and sidelights using custom YOLOv5 model with 95% mAP accuracy.

Python YOLOv5 OpenCV
Google Play Store App Analysis
Data Analysis

Google Play Store App Analysis

Comprehensive analysis of Google Play Store apps using Python and data visualization libraries.

Python Pandas Matplotlib
FootBalance Nepal Website
Web Development

FootBalance Nepal Website

Professional e-commerce website for FootBalance Nepal footwear company.

HTML CSS JavaScript
Nobel Prize Winners Analysis
Data Analysis

Nobel Prize Winners Analysis

Statistical analysis of Nobel Prize winners throughout history.

Python Pandas Data Visualization
Blog Webapp with User Authentication
Web Development

Blog Webapp with User Authentication

Full-stack blog application with secure user authentication and commenting system.

Flask Python SQLAlchemy
Flight Finder Bot
Backend Development

Flight Finder Bot

Automated flight price tracker with notification system.

Python API Integration Tequila API

Tableau Dashboards

Customer Churn Analysis

Click to View

Customer Churn Analysis

Modeling Frequency of network use and customer complains for an Iranian Telecom Company.

Call and Tariff performance

Click to View

Call and Tariff performance

Modeling call failures and evaluating tariff plan preference for an Iranian Telecom Company.

Achievements

Contact Me

I'm always open to discussing research opportunities, collaborations, or exciting projects. Feel free to reach out!

Location

Edwardsville, Illinois