Project Portfolio

Project
Showcase

Bridging the gap between robust software engineering and cutting-edge artificial intelligence.

Project Showcase

Project Type

🔧Filter by Technical Stack

Select any technology to view projects using that stack

Languages
Python
TypeScript
C#
Frontend
Next.js
React
Tailwind CSS
Framer Motion
Blazor
Backend
Node.js
Supabase
PostgreSQL
Serverless Functions
ASP.NET Core 8.0
Azure SQL
AI/ML
PyTorch
Computer Vision
Faster R-CNN
Hugging Face
AI Search Algorithms
NLP
Machine Learning
RAG
LangChain
AWS Bedrock
Azure OpenAI
BERT
VADER
Cloud/DevOps
Vercel
GitHub Actions
AWS Lambda
Azure
Azure Functions
APIs/Security
RESTful APIs
Resend API
SOC 2
Row-Level Security
HIPAA
Stripe
Medina Grill Official Website
Client Success

Medina Grill Official Website

High-performance restaurant website featuring mobile-first design, automated catering lead generation, and seamless integration with delivery platforms.

Next.js
Tailwind CSS
Framer Motion
Vercel Edge
Resend API
Resume Database Web Application
Client Success

Resume Database Web Application

Production-ready candidate management system with enterprise security, SOC 2 compliance, and real-time analytics for recruitment tracking.

Next.js
Supabase
PostgreSQL
TypeScript
SOC 2
Row-Level Security
AI Pacman Maze Solver
Academic Research

AI Pacman Maze Solver

Implementing classic AI search algorithms (DFS, BFS, A*) to solve complex Pacman maze navigation problems with heuristic optimization.

Python
AI Search Algorithms
Pathfinding
Heuristics
Orthopedic Screw Classification
Academic Research

Orthopedic Screw Classification

Medical imaging AI system using Mask R-CNN to detect and classify orthopedic screws from X-ray images. Trained on GPU-accelerated infrastructure with custom-labeled dataset of 2,000+ X-ray images. Achieves instance segmentation and multi-class classification for surgical hardware identification.

Python
Mask R-CNN
PyTorch
GPU Training
Medical Imaging
Instance Segmentation
Computer Vision
Pedestrian Detection with Faster R-CNN
Academic Research

Pedestrian Detection with Faster R-CNN

Deep learning-based pedestrian detection using Faster R-CNN architecture for autonomous safety and urban surveillance applications.

PyTorch
Computer Vision
Faster R-CNN
Deep Learning
Reddit Sentiment Analysis Engine
Academic Research

Reddit Sentiment Analysis Engine

NLP-powered sentiment analysis tool using VADER lexicon for analyzing emotional tones in social media content, deployed on Hugging Face Spaces.

NLP
Sentiment Analysis
VADER
Hugging Face
Image Feature Matching Lab
Academic Research

Image Feature Matching Lab

Computer vision implementation comparing ORB, SIFT, and FLANN algorithms for feature detection and matching in image processing tasks.

OpenCV
Computer Vision
Feature Matching
SIFT
Object Detection with Faster R-CNN
Academic Research

Object Detection with Faster R-CNN

Interactive web interface for real-time object detection using Faster R-CNN model trained on COCO dataset, capable of identifying 80+ object classes with adjustable confidence thresholds.

PyTorch
Computer Vision
Faster R-CNN
Hugging Face Spaces
ITSE Full-Stack Vercel Deployment
Client Success

ITSE Full-Stack Vercel Deployment

An end-to-end deployment showcase highlighting modern cloud architecture, serverless workflows, and automated CI/CD pipelines for production-ready SaaS applications.

Next.js
Node.js
Vercel
GitHub Actions
Serverless Functions
Tailwind CSS
AI-Powered Educational Gaming & Analytics Platform
Academic Research

AI-Powered Educational Gaming & Analytics Platform

Research project at Saint Louis University developing an intelligent gamification system that uses machine learning to analyze student engagement patterns and optimize educational outcomes through adaptive game mechanics.

Python
Machine Learning
Educational AI
Data Analytics
Gamification
AI Document Analyzer
Academic Research

AI Document Analyzer

A web app that lets users upload PDF documents and ask questions about them using AI. The system reads the PDF, converts it to searchable embeddings, and uses Google's Gemini AI to answer questions based only on the document content.

Flask
Python
LangChain
Chroma
HuggingFace
Google Gemini
JavaScript
HTML/CSS
RAG
Vector Database
Arogya Vidya - Full Stack Healthcare Platform
Client Success

Arogya Vidya - Full Stack Healthcare Platform

Comprehensive healthcare management platform built on Azure featuring patient records, appointment scheduling, telemedicine integration, and AI-powered health insights for modern medical practices.

Next.js
React
Node.js
Azure SQL
Prisma
TypeScript
Tailwind CSS
Stripe
AI/ML
Azure
FlavorBot - Restaurant RAG Assistant
Client Success

FlavorBot - Restaurant RAG Assistant

Production-grade RAG system deployed on AWS serverless architecture for restaurant operations. Uses AWS Bedrock, Lambda, OpenSearch for intelligent menu recommendations and 24/7 customer support. Handles 500+ daily queries with 97% accuracy and 99.9% uptime.

AWS Lambda
AWS Bedrock
OpenSearch
RAG
LangChain
S3
API Gateway
Next.js
CloudWatch
SupportAI - Enterprise IT Support RAG
Client Success

SupportAI - Enterprise IT Support RAG

Enterprise RAG system on Microsoft Azure transforming IT support operations. Leverages Azure OpenAI, Cognitive Search, and Functions for expert troubleshooting. Achieves 94% accuracy, 60% reduction in L1 escalations, and SOC 2 compliance with $340K annual savings.

Azure OpenAI
Azure Functions
Cognitive Search
RAG
Cosmos DB
TypeScript
Service Bus
LangChain
GovSecure - State Government Citizen Portal
Client Success

GovSecure - State Government Citizen Portal

Production-ready .NET 8 enterprise platform for state government agencies built on Azure. HIPAA and SOC 2 Type II compliant system enabling secure citizen data collection with encrypted database storage and mTLS government integration. Features AES-256 encryption, immutable audit logs, and 99.95% uptime SLA.

ASP.NET Core 8.0
C#
Azure SQL
Azure AD B2C
Blazor
Entity Framework
HIPAA
SOC 2
Azure
Reddit Sentiment Analysis Pipeline
Academic Research

Reddit Sentiment Analysis Pipeline

Production-grade ML pipeline extracting Reddit threads and performing real-time sentiment analysis using VADER and BERT transformers. Processes 10K comments/hour with 89% accuracy. Features interactive D3.js visualizations, keyword filtering, time-series trends, and CSV/JSON export capabilities.

Python
BERT
PyTorch
VADER
Apache Kafka
PostgreSQL
D3.js
Next.js
NLP
Machine Learning
Medina Grill Merchandise Store
Client Success

Medina Grill Merchandise Store

Full-stack e-commerce platform built with React and Azure enabling branded merchandise sales. Features Stripe payment processing with 98.7% success rate, automated order management, and Resend email notifications. Includes admin dashboard, inventory tracking, and real-time analytics generating $45K+ monthly revenue.

React 18
Next.js
TypeScript
Stripe
Azure SQL
Resend API
Node.js
Tailwind CSS
Flower Recognition AI
Academic Research

Flower Recognition AI

Lightweight YOLOv8-powered image classifier achieving 98.6% accuracy in identifying 5 flower species. Built in under 100 lines of Python with complete beginner-friendly tutorial. Processes images in <4ms using transfer learning and deployed on Hugging Face Spaces with Gradio interface.

Python
YOLOv8
Computer Vision
Deep Learning
Gradio
Hugging Face
Transfer Learning
Machine Learning