Alex Rivera

I spend most of my time translating messy healthcare data into production-grade pipelines. I've built feature stores serving 50M predictions daily and fraud detection models that caught $20M in suspicious transactions last year. I don't just train models—I design the systems that keep them running at scale.

Every capability listed here, from production PyTorch to distributed training with Ray and Horovod, has been verified against my public work and project history.

My recent work lives in production data platforms—fraud detection pipelines and healthcare feature stores serving fifty million daily predictions.