Welcome

Hi, I'm Hardip Singh Dhaliwal — a CompTIA Security+ certified Computer Science professional and IEEE-published researcher based in Fresno, California, currently pursuing my M.S. in Computer Science at CSU Fresno.

I work across software development, process automation, and machine learning — from building Chrome extensions and Electron apps to optimizing neural networks. I like building efficient, practical tools that solve real problems.

CompTIA Security+ Machine Learning Process Automation Chrome Extensions Neural Networks Web Development
Hardip Singh Dhaliwal

Research & Publications

IEEE SMC 2024 — Kuching, Malaysia · October 2024

Post-Training Quantization in Brain-Computer Interfaces Based on Event-Related Potential Detection

H. Cecotti · D. Dhaliwal · H. Singh · Y. Meena

Evaluation of post-training quantization (PTQ) on state-of-the-art brain-computer interface models for single-trial detection of event-related potentials, enabling lightweight, fully portable BCI systems.

15× Model size reduction
0.825 AUC retained (from 0.861)
Post-Training Quantization Brain-Computer Interfaces ERP Detection
Read on arXiv →

CSCSU 2024 Symposium — California State University · Spring 2024

Weight Quantization in Extreme Learning Machines: Application to Handwritten Character Recognition

H. Cecotti · H. Singh — USDA–CSU Fresno Partnership Program

Research on compressing multilayer perceptron (MLP) neural networks through weight pruning and quantization, enabling efficient deployment in resource-constrained environments.

88.5% Model size reduction
97% Accuracy retained
5 Datasets evaluated
  • MATLAB pipeline for training, pruning, and quantizing MLP weights at 4, 8, and 16-bit precision
  • Evaluated on Bangla, CVL, Devanagari, MNIST, and Oriya handwritten character datasets
MATLAB Neural Networks Weight Quantization Model Compression

Projects

AI Job Application Filler

JavaScript Chrome Extension (MV3) Gmail API AI APIs

Chrome extension that auto-fills job applications across ATS platforms using AI-powered field mapping.

  • Supports Workday, iCIMS, Greenhouse, and Paycom with field mapping via Claude, Gemini, and Groq APIs
  • Gmail OAuth integration for AI-generated cover letters (A/B tested) with tracker dashboard and CSV export
  • Easy Apply automation for LinkedIn/Indeed with multi-step form navigation and confirmation-gated submission
View on GitHub →

EverView

Electron JavaScript Node.js

Cross-platform desktop app for floating, always-on-top PDF viewing with a built-in virtual keyboard.

  • BrowserView embedding with IPC communication and always-on-top window management
  • Virtual keyboard with QWERTY and symbol modes, opacity control, and find-in-page
  • Packaged installers for macOS (Apple Silicon) and Windows
View on GitHub →

Raspberry Pi Homelab

Linux WireGuard Pi-hole systemd

Self-hosted home server on a Raspberry Pi 5 — including the very site you're reading right now.

  • PiVPN (WireGuard) for encrypted remote access to the home network
  • Pi-hole for network-wide DNS ad and tracker blocking
  • Self-hosted web server running as persistent systemd services

Face Recognition with CNN

Python MATLAB OpenCV

Convolutional neural network for face recognition trained on the LFW (Labeled Faces in the Wild) dataset.

  • Data preprocessing pipeline with OpenCV and scikit-learn
  • Evaluated with confusion matrices, ROC curves, and multi-threshold accuracy benchmarks
View on GitHub →

AI Path-finding Maze Game

C++ Dijkstra's Algorithm Data Structures

Survival maze game where enemies use Dijkstra's algorithm to intelligently track and chase the player.

  • Dynamic enemy pathfinding with custom data structures
  • Interactive game mechanics with multiple levels of increasing difficulty
View on GitHub →

Online Checkers Game

MySQL PHP JavaScript

Multiplayer checkers game with user authentication and a competitive leaderboard system.

  • Secure login and registration with MySQL-backed user data and rankings
  • Real-time game mechanics and competitive leaderboard with player statistics
View on GitHub →

Instagram Unfollowers

JavaScript Chrome Extension

Chrome extension that finds and helps unfollow Instagram accounts that don't follow you back.

View on GitHub →

Netflix Data Analysis System

SQL Database Design Data Analysis

Comprehensive database system for analyzing Netflix content and user preferences.

  • Normalized schema with complex SQL queries for content analysis
  • Content recommendation and trend analysis systems