Caressa Arias
Say hello →

Plugin-based GIS platform for field sensor integration

Software Engineer (Secret Clearance) · Chesapeake Technology International · 2014–2018

Problem

The product was a Java-based GIS platform with a plugin architecture, and each government contract brought a different set of military hardware and sensors that had to appear on the map. The hardware talked over whatever it talked over — Bluetooth/BLE, serial, WiFi — with bandwidth and reliability that varied by device and by deployment. Every new sponsor requirement was therefore an integration problem before it was a visualization problem.

My role

I built custom plugins against the platform’s APIs to meet sponsor requirements, contributed to a lightweight AngularJS web GIS interface, and mentored interns through project design, daily check-ins, and code reviews. I also talked directly with sponsors and stakeholders to clarify requirements and demonstrate functionality, and took part in PI-style multi-month planning cycles to scope contract-driven work.

What I owned

Plugins that ingest, process, and visualize real-time sensor data on GIS maps. Protocol adaptation across BLE, serial, and WiFi, accounting for varying bandwidth and reliability constraints. Data normalization and mapping logic so sensor output was represented accurately in the GIS environment. Selenium-based automated testing support for the web components, and build maintenance and continuous integration through Jenkins. I also participated in on-site exercises to support deployed systems, troubleshoot issues, and validate software behavior in real-world conditions.

Key decision

I normalized every sensor’s output into one internal mapping model at the plugin boundary, rather than letting each integration hand its own shape to the GIS layer. That put the translation cost in the plugin — one adapter per device, written once — and kept the mapping and visualization code protocol-agnostic, so the next contract’s hardware did not require touching the display path. It does mean a sensor with genuinely unusual semantics has to be flattened into the common model, which is a real loss of fidelity in exchange for a platform that stays integrable contract after contract.

Result

I delivered contract-specific sensor integrations across multiple deployments and supported those systems in the field during exercises. Alongside the integration work I mentored four interns through design, review, and daily guidance.

Built with: Java · Android · AngularJS · BLE / serial / WiFi · Jenkins · Selenium