Data Scientists Should Build Dashboards (and write ETL too)
Bringing machine learning (ML) powered products to a large production environment requires a diverse skillset: ML experts who can identify which approach/algorithm will best fit the problem at hand, engineers who can take the prototypes from the ML experts and productionize them, and analysts who can measure how successful the solution has been. It is common to see companies organize themselves into teams that are responsible for each of these tasks: data scientists, data engineers and data analysts respectively. At Shopify we believe this is sub-optimal, as understanding of the actual problem being solved leaks at each interface. Instead we build data science teams that are responsible for all three of these tasks. This talk will outline how we organize our data science teams to meet these challenges and bring ML powered products to production.
Kyle Tate is a senior data science lead at Shopify. He leads the teams responsible for building the data products that power Shopify Capital, Fraud Protect for Shopify Payments and Shopify's internal risk management. Kyle is interested in taking the ideas and techniques from the pages of machine learning textbooks and papers and bringing them to life in the real world.