Productionising ML models in an enterprise environment

This session will talk about Wesfarmers OneDigital’s MLOps journey in taking a machine learning model from the proof of concept stage to the production stage. This session will cover different aspects of model productionisation around people, process, and technology in an enterprise environment and present the approach at tackling these challenges at Wesfarmers.

Key Takeaways

• Define what productionisation means in your situation
• An evaluation of business value vs risk can be applied to models
• Consider treating machine learning models as you would other products
• RASCI Matrix is simple but very useful
• You may need to prioritise things like security, privacy and cost management

————————————————————————————————————————————————————

Bio

Paridhi M – Senior Machine Learning Engineer | Wesfarmers OneDigital | Australia

Paridhi is an experienced professional with over 14 years of experience in machine learning and software engineering. She has proven experience in building and deploying end-to-end statistical and machine learning models in large enterprise environments in e-commerce (ads classification, fraud detection and recommender systems), retail (facial recognition and age/gender classification) and banking (faulty payments, sentiment analysis). Her strengths lie in software engineering, machine learning engineering, ML production systems and emerging MLOps technologies. She holds a Master’s degree in Machine Learning from University of Edinburgh and a Bachelor’s degree in Computer Science from National Institute of Technology, Allahabad.