Computer Vision / Railway AI / Applied Engineering
Building practical AI systems around railway inspection, vision models, GPS context, depth sensing, reporting workflows, and deployment utilities.
Profile
My work sits around applied AI, computer vision, geospatial context, and hardware-aware software. The current flagship project is a real-time railway pantograph monitoring system using camera feeds, ZED depth data, GPS files, calibrated measurement logic, alerts, and generated reports.
This portfolio is published with GitHub Pages and links directly into the public code, reports, maps, and project repositories.
Selected Work
A focused set of public projects across inspection AI, accessibility tooling, intelligent transport logic, and knowledge mapping.
A production-oriented inspection workflow combining YOLO detection, ZED depth processing, GPS route context, measurement formulas, alert logging, and report generation.
Accessibility-first captioning project for visual impairment assistance, built around live speech and readable assistance workflows.
A fuzzy-logic highway project exploring decision support around safe joining speed and intelligent traffic behavior.
A Reddit scraper project for turning saved posts into a more navigable mind-map style knowledge base.
Case Study
The railway project uses separate producer and consumer processes, shared memory frame transfer, calibrated pixel-to-millimeter conversion, GPS mast matching, ZED depth aggregation, and FFmpeg video writing.
Pantograph, mast, contact point, and arm detection workflows.
Height, stagger, gradient, double contact, and setting distance.
Alerts, maps, reports, launchers, camera checks, and conversion tools.
Stack
The portfolio keeps heavy model binaries out of GitHub while preserving the source code, documentation, maps, and utility scripts needed to understand the engineering work.
Contact
The GitHub profile is the main launchpad for public repositories and future project updates.