Machine learning / MLOps / AI automation

Models don't ship themselves.
So I build the pipeline that does.

I train the model, wrap it in a pipeline that retrains and redeploys it, and automate everything either side of that.

I'm an AI/ML engineer working across the whole lifecycle: data collection, training, and the MLOps around it — containerised services, scheduled retraining, monitored endpoints. I learned it from underneath, writing classifiers in raw NumPy and file systems in C before trusting a framework, which is why I can tell a bad gradient from a bad deploy. Most of what I build now is automation: the scraper that feeds the model, the job that retrains it, the API that serves it.