Abstract: 

Whether you are a veteran Data Science practitioner, a novice ML engineer, or a hard-working DevOps ninja, you probably heard about MLOps. But what are MLOPs? And do they only relate to applied machine learning in production?

Suppose the proverbial google query “what is/are MLOps” was done recently. In that case, you might have discovered that while most MLOps involve some “deployment” or “production” elements, an MLOps strategy affects data-related projects from square one. You will also find many different blueprints, suggested roles, technology stacks, etc. Often, the meaning of MLOps depends on who is asking and who is answering. There are many cases where several simultaneous working definitions of MLOps exist even in the same organization. Rightfully, there are fears that MLOPs changed from being an engineering terminology to marketing jargon.

So, what does the term MLOps mean today?

Join me as we deconstruct several definitions of MLOps and find the essential practices which even a solo researcher should adopt. We will distill the demands for reproducibility and deployments to a bare minimum shared by all stakeholders in the operational ML process. Searching for this “lowest common denominator,” we will discover the fundamental charter of MLOPs, and draft basic recipes for incrementally crafting functional MLOps units - already during R&D.

Bio: 

Researcher first, developer second. Over the last 5 years, Ariel has worked on various projects; from the realms of quantum chemistry, massively-parallel supercomputing to deep-learning computer-vision. With AllegroAi, he helped build an open-source R&D platform (Allegro Trains), and later went on to lead a data-first transition for a revolutionary nanochemistry startup (StoreDot). Answering his calling to spread the word, he recently took up the mantle of Evangelist at ClearML. Ariel received his PhD in Chemistry in 2014 from the Weizmann Institute of Science. Ariel recently made the transition to the bustling startup scene of Tel-Aviv, and to cutting-edge Deep Learning research.