My name is Vin. I help companies make machine learning work in real life. I’ve worked at companies like eBay, IBM and Samsung.

I build end-to-end, cloud native ML systems. In practice, that usually means a combination of models (TensorFlow and PyTorch) and data pipelines (Beam, Pub/Sub, Luigi, Spark) served over Google Cloud.

In my free time, I enjoy reimplementing popular packages/libraries (eg PyTorch, Luigi) from scratch, annotating seminal papers in machine learning, creating impractical tools (eg remote code execution tools, classification tasks using evolutionary strategies, serverless mapreduce), dissecting software (eg Seaduke, Lisp interpreters) and contributing to open source projects (sparse-dot, scikit-network, tsfresh, scikit-sound, Textblob, Jazzband).