RuralReach
DNN
Contextual AI
This project was all about making AI understand context. Me and my friend built a deep learning system that doesn't just look at raw data, but also considers the "why" and "where", like environmental and situational cues, to predict behaviors more accurately.
We handled the entire pipeline ourselves, from engineering the features to training the neural networks, making sure the model stayed sharp even when the situation changed.
NBA Match Tracker
Data Visualization
API Integration
Tired of clunky sports sites, I built this dashboard to be a snappy, minimal way to check NBA scores.
It hits the balldontlie API for live game data and uses Recharts to turn raw box scores into visual trends.
It was a great excuse to play with a high-performance React/Vite stack while keeping tabs on the league, complete with glassy UI cards and responsive filters for the ultimate fan experience.
View Project
Semantic Talent Indexer
NLP
Ollama
Semantic Search
This is my take on making hiring smarter and more secure. I built a RAG (retrieval-augmented generation) system that uses semantic search to find the best talent, moving way beyond basic keyword matching. I did the work of setting up a local inference pipeline using Ollama so that all the sensitive candidate data stays private and on-premises.
It was a huge win in showing how we can use large language models (LLMs) to understand the actual context of someone's career, not just their job titles.
View Project
HealthPredict
Healthcare AI
Patient Care
Predictive Analytics
This project uses machine learning to predict if a patient is likely to miss their medical appointment. I built a random forest model to spot potential "no-shows," helping doctors optimize their schedules and make sure more people actually get the care they need.
I did the work of balancing the data and prioritizing recall because, at the end of the day, missing a patient who needs help is a bigger deal than a false alarm.
View Project
Weather Analysis
Exploratory Data Analysis
This project started with a massive, messy pile of atmospheric data and a lot of questions. I built a Python pipeline to clean everything up and get it ready for some serious time-series analysis. I visualized using Matplotlib and Seaborn to see if I could spot seasonal trends or weird atmospheric shifts.
It was basically me playing detective with data to find patterns and anomalies that aren't obvious at first glance.
View Project
Explain
Artificial Intelligence
EdTech
Data Visualization
Math shouldn't be intimidating. I built Explain to act as a universal translator for complex science, taking "scary" formulas like Navier-Stokes or Schrödinger and breaking them into plain English.
It uses a high-contrast minimalist interface where AI-driven explanations meet vibrant, animated visualizations, turning abstract symbols into something you can actually see and understand.
View Project
Student Application
JSON
REST API
This is a full-stack student management system I built to bridge the gap between backend and frontend. I wanted to show how a Spring Boot API and a React UI can work together perfectly on a single port.
I did the whole thing from scratch, designing the RESTful architecture, building the component-based interface, and making sure the state management was smooth for every add, update, and delete.
View Frontend Code
View Backend Code
B+-Tree Index Manager
File I/O
Memory Management
Data Structures
This was probably one of my biggest coding challenges yet. I built a fully functional b+-tree index manager in c to help a database find and retrieve records without breaking a sweat. I did the heavy lifting of mapping tree nodes to disk pages and implementing all the core logic, like splits, merges, and range scans.
I also integrated it with a custom buffer manager to make sure the whole system was fast, efficient, and actually persistent.
Record Manager
File I/O
Memory Management
Data Structures
I wanted to dive deep into the guts of a database. I built a record manager in C from the ground up to handle data tables with fixed schemas. I did the hard work of implementing the core logic, like how records are inserted, deleted, and scanned, using my own custom data structures.
I designed the page layouts and integrated a buffer manager to make sure accessing data was as fast and efficient as possible.
