Speaker "Maxim Fateev" Details Back



Manage Full ML Model Lifecycle with Temporal


Designing and training is just one step on the road to production for an ML model. A complete solution should encompass the full lifecycle including production deployment, performance monitoring and retraining. Every organization has its own needs and processes, thus it is commonplace to invest time in building reliable custom solutions from existing open source components. Existing service orchestration frameworks do not make this task simple as none can be used to fully automate the process end to end.
Temporal is an open source service orchestration platform that is used by multiple companies to automate their core ML, Big Data and CI/CD pipelines end to end. The talk gives an overview of Temporal and discusses architectural and design patterns applicable to ML use cases.


Maxim has over 20 years of experience building large scale distributed systems at Amazon, Microsoft, Google and Uber. While at Amazon, he led the design and development of Amazon Messaging System that later was adopted as the AWS SQL backend. Afterwards,  he was a tech lead of the AWS Simple Workflow Service. At Uber, Maxim led both the Cherami open source messaging project and later Cadence Workflow. In October of 2019, Maxim cofounded Temporal Technologies where he has led as CEO since.