Software engineer and systems architect focused on reliable research tools.

My background spans three decades of financial-services technology, from software development and team leadership to systems architecture. I spent roughly 15 years working in Java. Since 2017, Python has been the focus of my private project work.

I'm now bringing that experience to research software, data pipelines and applied ML, with a particular interest in cancer and biomedical research. The work that appeals to me is practical: make the inputs trustworthy, the processing reproducible and the results easier to inspect.

I'm open to conversations about remote, part-time software and data-engineering roles with research teams. I'm especially interested in work where dependable tooling can save researchers time: integrations, pipeline reliability, validators, testing and model-evaluation workflows.

• PREDVAL — A predictions-only validation harness for examining model coverage, calibration and cluster-aware uncertainty. Further work on patient-partition checks, convergence handling and reproducible comparisons is available in a draft pull request.

• Histopathologic Cancer Detection — A config-driven TensorFlow/Keras project for PatchCamelyon, with whole-slide-image grouping to reduce train/validation leakage. Includes training, evaluation and prediction workflows.

• DataHoover — Data ingestion and automation with an emphasis on provenance and repeatable runs. Recent reliability work adds shared warehouse locking and idempotent ingestion.

PREDVAL and Histopath are research projects, not clinically validated products. The repositories document methods, checks and limitations.

My foundation is systems architecture and integration. A central part of my architecture work is deciding which component should own each responsibility, with clear service boundaries and manageable dependencies. That includes understanding where data comes from, what an interface promises and how failures propagate. I value clear requirements, reproducible tests, explicit limitations and review before changes are accepted.

I use AI tools as part of my development workflow. I'm interested in teams that combine that capability with clear data-handling rules, testing and accountable review.

Current focus: Python, data pipelines, automation and ML-evaluation tooling.
Earlier professional experience: Java/Spring, service integration and large financial systems, with extensive web-services experience (primarily SOAP, with some REST).

For technical discussion, open an issue in the relevant repository or tag @aRealGem in a thread.

Outside code, I'm a Taekwondo black belt and a spaceflight-simulator enthusiast.