01 /

Projects & Systems

Transportation

Dynamic Transportation Network Modeling

Essential tool for transportation planners and policymakers, providing insights to optimize system performance. Large-scale network simulation for urban congestion analysis and infrastructure planning.

Network Simulation Urban Planning

Real-Time Traffic Prediction

Deep learning models predicting traffic speed, anomalies and incidents 30 minutes ahead across large highway networks, enabling proactive traffic management. Deployed with live data pipelines for transportation agencies.

Machine Learning GRU Models Probe Vehicle Data

Large-Scale GPS Trajectory Processing

Map matching algorithms over 4+ TB of GPS probe vehicle data, deriving stops, tours and route choices for individual vehicles. Forms the trajectory layer that downstream demand and route-choice analysis is built on.

Map Matching 4+ TB Probe Vehicle Data

Real-Time Signal Timing Recommendation

Signal timing plan recommendations generated in real time, combining plans tuned for possible incidents with historical patterns so intersections adapt to conditions as they develop rather than after the fact.

Signal Control Real-Time Incident Response

COVID-19 Mobility Impact Analysis

Assessed pandemic impact on downtown Pittsburgh traffic by visualizing speed and count data before, during, and after COVID-19. Interactive web dashboards for policy makers.

Data Visualization Web Development

Flexible Mobility Services

Developing solutions for flexible travel options with expanded riding hours and dynamic pick-up/drop-off locations. Deployed the Rural Access Mobility Platform (RAMP) in Greene County, PA to improve transportation access in rural communities.

Mobility Services RAMP Rural Access

Sensor-Less Parking

Real-time spot-occupancy management in managed lots without installing per-spot hardware sensors — state is derived from a JWT-authenticated reservation and check-in flow, with automatic spot allocation and email reminders for upcoming bookings. Implements the non-sensor parking concept of Yang & Qian (CMU CEE) on a Django REST + React stack; deployed at the CMU Morewood lot and structured to scale.

Sensor-Less Occupancy Reservation Flow CMU Morewood

Landslide Early Warning System (Maryland DOT)

Developed precipitation-based landslide risk assessment and alert dashboard for Maryland State Highway Administration. Integrates real-time weather data, soil moisture prediction, and geospatial slope analysis for statewide monitoring and early warning.

Risk Assessment Open-Meteo API GIS Integration
Physics

High-Energy Physics (ATLAS, LHCb)

Contributed to major experiments at CERN including Higgs boson searches, CP violation measurements, and detector development. Published in Physical Review Letters and Physical Review D.

CERN Particle Physics

CVD Diamond Detector Development (RD42)

Designed and built pad, strip, pixel, and 3D CVD diamond detectors for extreme radiation environments. Integrated Diamond Beam Monitors into ATLAS readout systems.

Detector Physics Instrumentation