MLOps as Tool to Shape Team and Culture

Good MLOps practices are a way to operationalize a more “vertical” practice and blur the boundaries between different stages of “production-ready”. Sometimes you have this idea that production-ready means global availability but with ML products that need to be constantly tested against real-world data, we believe production-ready should be a continuum and that the key person that drives that needs to be the data scientist or the ML engineer.

Take-aways

In the last year or so, Ciro had to approach MLOps from the perspective of an organization that needs to grow efficiently. The ML team at Coveo more than doubled in a few months, raising several issues about productivity, retaining talent, and honestly even just making sense of the org chart. These are all happy problems because they’re growing, but problems nonetheless. They all have to do with a better developer experience and stronger experimentation culture. Also, some of these concepts might seem obvious but - believe Ciro - they often need to be evangelized a lot. Like a lot! For a company that has essential ML components in the product (like Coveo which sells search and recommendation APIs) the ultimate goal is to have a bunch of ML capabilities embedded in the product. To do that your ML engineers need to be trained in the business problem as much as they are in hyperparameter optimization. So having 20 people that do ML as a separate unit in your company might not be the right way to go. On the other hand, if the business units are not prepared to absorb the ML engineers with the right infrastructure, tooling, and processes they will not make an impact. Here is where MLOps became an active area of interest. Ciro works with Jacopo Tagliabue and together they tried to approach this problem also from an MLOps and developer experience perspective.

In this episode

Ciro Greco

VP of AI, Coveo

Ciro Greco, VP of AI at Coveo. Ph.D. in Linguistics and Cognitive Neuroscience at Milano-Bicocca. Ciro worked as visiting scholar at MIT and as a post-doctoral fellow at Ghent University. In 2017, Ciro founded Tooso.ai, a San Francisco-based startup specializing in Information Retrieval and Natural Language Processing. Tooso was acquired by Coveo in 2019. Since then Ciro has been helping Coveo with DataOps and MLOps throughout the turbulent road to IPO.

Demetrios Brinkmann

Host

Demetrios is one of the main organizers of the MLOps community and currently resides in a small town outside Frankfurt, Germany. He is an avid traveller who taught English as a second language to see the world and learn about new cultures. Demetrios fell into the Machine Learning Operations world, and since, has interviewed the leading names around MLOps, Data Science, and ML. Since diving into the nitty-gritty of Machine Learning Operations he felt a strong calling to explore the ethical issues surrounding ML. When he is not conducting interviews you can find him making stone stacking with his daughter in the woods or playing the ukulele by the campfire.

Vishnu Rachakonda

Host

Vishnu Rachakonda is the operations lead for the MLOps Community and co-hosts the MLOps Coffee Sessions podcast. He is a machine learning engineer at Tesseract Health, a 4Catalyzer company focused on retinal imaging. In this role, he builds machine learning models for clinical workflow augmentation and diagnostics in on-device and cloud use cases. Since studying bioengineering at Penn, Vishnu has been actively working in the fields of computational biomedicine and MLOps. In his spare time, Vishnu enjoys suspending all logic to watch Indian action movies, playing chess, and writing.

Take-aways

In this episode

Ciro Greco

Demetrios Brinkmann

Vishnu Rachakonda

Upcoming episodes

Argo Workflows

Kemal Tugrul Yesilbek

FLOps with Scaleout's Open-core Platform

Marco Capuccini