AI Engineer | Energy & Mobility Systems
MS in Electrical & Computer Engineering from UNLV. Building intelligent systems for sustainable transportation and energy management through AI and data science.
Programming & Tools: Python, SQL, MATLAB, Git, Linux, Docker
Machine Learning & AI: PyTorch, TensorFlow, Scikit-learn, Deep Learning, Neural Networks, Graph Neural Networks
Data Science & Analytics: Pandas, NumPy, Jupyter, GIS Tools, Data Visualization, Statistical Analysis
Transportation infrastructure generates massive amounts of data, but turning sensor feeds into actionable insights for congestion, safety, and emissions remains a persistent challenge. I build open platforms and AI models that bridge research and operations—tools that work with messy, real-world data and deliver value to public agencies.
My journey started with computer science foundations and early work on large-scale digital infrastructure—civil registry systems, healthcare data platforms, and government service delivery—where I learned to design technology that scales to millions while meeting operational constraints like incomplete data, regulatory requirements, and 24/7 uptime across diverse stakeholders.
At the University of Nevada, Las Vegas, I shifted focus to U.S. transportation infrastructure through the USDOT REPS Tier 1 University Transportation Center, publishing in top IEEE venues:
IEEE ITSC 2024: "Using Deep Traffic Prediction for EMFAC Emission Estimation and Visualization"
IEEE ICVES 2024: "Benchmarking/Limitations of Traffic Prediction with Noisy Field Measurements"
My core innovation is graph neural networks with masked training that predict traffic patterns from noisy/missing sensor data—deployed across ~1,000 Nevada DOT freeway sensors.
Currently building Lattice, an open transportation analytics platform hosted at UNLV that provides real-time traffic prediction, corridor-level emissions estimation (EMFAC integration), interactive visualization, and REST API for agency use.
Selected projects built to be useful, measurable, and real.
Platform concept for real-time traffic operations combining graph neural networks with transformer forecasting. Includes KPI monitoring, congestion trend views, and short-horizon predictions for network-level decision support.
Real-time factory surveillance and automation using YOLO object detection, MediaPipe hand tracking, and multi-stream video processing. Deployed at Haig's Quality Printing for production monitoring and quality control.
Building production-grade autonomous navigation from scratch—differential drive kinematics, SLAM mapping, and PID control on Jetson Nano hardware running ROS Melodic. Educational robotics platform demonstrating embedded systems principles.
Simulating reusable launch vehicle descent with adaptive control strategies—RCS thrusters, hybrid aerodynamics, and grid fins across 40km atmospheric transition. 6-DOF dynamics with PID-based trajectory correction.
Competing in Sussex-Huawei Locomotion Challenge 2018—classifying walking, running, biking, and transit modes from smartphone IMU data using ensemble decision trees and wavelet features. Achieved 82.8% accuracy with resource-efficient algorithms.
Interested in collaborating on projects in energy engineering, machine learning, or sustainable transportation? Let's connect.