At a telecom company under NDA, I designed and built an automated fleet monitoring system for approximately 2.1K network devices. It cut a full collection cycle from approximately 24 hours to 2 hours and automated analysis and reporting that had previously been done manually.
I also replaced a sequential, partly manual device-processing script with a seven-stage Temporal pipeline that processes approximately 20K devices in around 10 minutes. I built an internal Python library for automating network devices across approximately 15 device types, and added support for six more device types to a production scheduling system.
On an independent contract project, I designed and delivered a production construction cost estimation platform in two months. It processes XLSX project files into estimates for accountant review, with overnight batch processing and supplier integrations.
At Multi-Agent Systems Laboratory, I built REST APIs and a web UI for configuring, launching, and monitoring ML training jobs. I also developed a SLAM module for optical navigation demos and co-authored four publications, including one Scopus-indexed publication.
