Go, ONNX and AI in Production – Andrew Wormald

Why watch

This talk presents a real-world case study of deploying machine learning models in a 24/7 automated trading company. The speaker covers the complete ML pipeline: Python for model training, ONNX for model portability across languages, and Go for high-performance live inference. It explores how classical ML, CNNs, and LLMs work together in the decision-making loop for trade execution and order sizing, with 90% of the production platform written in Go. Ideal for developers building production ML systems who want to understand practical patterns for integrating multiple model types into a unified automated system.