Projects

UT Dallas course project · January 2026 – May 2026

Led a 4-person team through three development iterations, assigned tasks, and presented the final demo. Won the course’s Best Implementation award.
Built the pipeline that turned a DFW Terminal D floor plan into a routable graph of 1,351 nodes, 5,944 edges, and 129 mapped amenities. Placed nodes over the plan image, georeferenced them from three known-coordinate control points, connected them with 8-directional edges, and exported the graph as generated SQL.
Built a tool that overlaid the graph on Google Maps so we could see and fix misplaced nodes.
Added routing to wheelchair-accessible and gender-specific restrooms, using PostgreSQL on AWS RDS behind the team’s 20+ endpoint Flask REST API.
Kept routes from crossing the TSA boundary by tagging each node as before or after security, so routing only uses nodes on the traveler’s side.
Developed an Android app in Kotlin using the Google Maps SDK with turn-by-turn guidance, off-path recalculation, and rerouting when amenities close or reach capacity.
Helped trace 36 functional requirements across 21 use cases and three IEEE-format test plans.
Served as class representative to faculty and helped the teams agree on shared requirements. Co-wrote the SRS, project plan, and design diagrams with my team.

GitHub: View Accessible Indoor Navigation System on GitHub

Personal project · March 2024 – Present

This site runs on a server at my home, not a hosting service.
It started as a Flask app on a Raspberry Pi with Gunicorn, Apache, SSL, and fail2ban.
I moved it to WordPress on an Ubuntu Server VM, first in VirtualBox and since October 2026 on Proxmox.
I split my home network into six VLANs on a managed switch so the web server sits on its own.

UT Dallas course project · August 2025 – December 2025

Built a regression pipeline on 3.1M+ Dallas Central Appraisal District records (2021 to 2025), reaching R-squared of 0.87 on a 30% held-out test set.
Built a six-stage Python ETL pipeline that automated web scraping, ZIP extraction, multi-table joins, categorical encoding, and data validation.
Integrated three government data sources (DCAD, Dallas Open Data, Texas Education Agency) using composite keys and pandas merge operations with configurable join strategies.
Performed feature selection using Pearson correlation for numeric variables and eta-squared for categorical variables, identifying ZIP code (η² ≈ 0.31) and school district (η² ≈ 0.22) as top predictors.
Trained and benchmarked five regression models (Linear Regression, Decision Tree, Random Forest, XGBoost, LightGBM), using GPU acceleration to cut gradient boosting training from 10+ minutes to under 15 seconds.

UT Dallas course project · January 2025 – May 2025

Led a 6-person team, turned sponsor interviews into Jira stories, and presented progress at weekly sponsor meetings.
Provisioned and secured an AWS EC2 GPU instance to host 7B to 14B parameter models, with role-based access so users could switch between them.
Set up and tested Ollama and Open WebUI to run the models.
Wrote a setup script that automated GPU detection, dependency installation, security patches, LUKS-encrypted Docker storage, and service configuration.

Personal project · May 2024 – September 2024

I built my own thermostat for my home’s HVAC.
I soldered the relays that drive the heating and cooling, wired the temperature sensors with their resistors so they read correctly, and ran it all on an ESP32 with Tasmota.
Home Assistant runs the automation rules, including fail-safes that never run heat and cooling at the same time and that prevent compressor short-cycling.
A second sensor in another room lets it run on the whole-house average.

UT Dallas course project · January 2026 – May 2026

Connected the team’s React front end to the Flask backend so the screen showed only device state the backend had saved, across seven device types.
Fixed corrupted device state when several changes landed at once by queuing writes to the state file behind a lock.
Built the event log, which saved each device change with its reason and showed the last 50 events in the dashboard.
Built the notification system tested in a 5-person usability study, one version per iteration: an activity log, Discord webhook alerts, then browser push notifications.
Ran all three prototype versions at the same time so each participant could test every version.
Co-designed the Likert survey on trust, transparency, and usability, and ran 2 of the 5 participant sessions.

GitHub: View Interactive Smart Home Dashboard on GitHub

UT Dallas course project · August 2024 – December 2024

Developed a Python search engine using Key Word in Context (KWIC) indexing with the Flask framework.
Stored and indexed over 100,000 entries in MySQL on AWS RDS.
Used multi-threading to handle multiple user requests at once.
Scraped and indexed content from over 150 websites.
Added autocomplete, clickable links, and results ranked by how often users opened them.
Worked with the team on requirements and architecture, and used GitHub for version control.
Wrote the documentation and system diagrams, and presented to stakeholders.

GitHub: View KWIC Web Search Engine on GitHub

Personal project · March 2024

I built a home security camera on a Raspberry Pi running MotionEyeOS.
It detects motion and saves video to a separate Ubuntu Linux server I set up, over SSH.

UT Dallas course projects · 2024

Indoor Navigation App Requirements: stakeholder analysis and requirements like real-time audio feedback, hazard detection, and route optimization.
Database and Webpage Project: website built with HTML, Bootstrap, Flask, and SQLAlchemy. View on GitHub
Mock Google Budget Integration: schedule, budget, and risk plan in Microsoft Project.
Doctor’s Office Multithreading Simulation: threads for patients, receptionist, nurses, and doctors.
CPU and Memory Simulation: two processes talking through pipes and forks, built on Linux.