Last week (7th-10th of June 2022) we had the pleasure to travel all the way to Toronto together for the MLOps world conference 2022!
It was a 4 day conference of which two days were online and filled with workshops and talks, and two days were in-person at the Sheraton Conference center. It was really nice to be able to finally go to an in-person conference again and see our MLOps community friends in real life.
The conference focused on anything related to productionizing Machine Learning, a perfect fit for us. We had the privilege of having two talks, one about tackling zero downtime in ML, and one about deploying and orchestrating with UbiOps, as well as a booth in the main hall. Recordings should become available soon if you missed the conference!
The majority of the attendees were practitioners such as data scientists and machine learning engineers, but also C-level and company owners.
Fellow software scale-ups from different places all over the world were present and it was great to finally meet them. For example, our friends at Arize (California), but also the teams of Piscellia (France) and Aporia (Israel) were there.
The MLOps world conference is the first conference ever that is solely focused on MLOps. As the industry matures, more companies will encounter this challenge one way or the other. We’re lucky to be among the first solution providers in the space and participate in this event. All the major players in the MLops space were there, as well as nearly everyone that we talked with was looking for a solution to operationalize their models securely, reliably, and cost-efficiently. They knew that building the stack themselves would cost a lot of time and money. If there is any conference that matches the industry UbiOps is in, it’d be this one.
The conference was a great success and we had a blast.
Wouter, Partnership Manager, had taken the initiative to organize the MLOps community drinks on the evening before the in-person part of the conference started. This definitely became the number one highlight of the week. So many people joined and we finally had the opportunity to meet Demetrios Brinkman in person! The drinks were an opportunity to learn more about what everyone else was doing in terms of MLOps, from their own MLOps companies to the use cases, they see as practitioners. The community is very vibrant and we would encourage anyone in the field of MLOps to join it.
The conference had much more to offer than just meeting fellow practitioners of course. We also attended talks in between manning our booth. A talk that really stood out to us was “Solving MLOps from first principles” by Dean Pleban. He accurately described how hard it is as a data science team to pick an MLOps tool that is right for you. Something we see among our users continuously. There are so many tools out there nowadays that it can be easy to get decision fatigue, or as Dean so aptly put it “MLOps fatigue”. He emphasized the importance of identifying your specific needs and finding the tool that best fits those, instead of just going for the hot new tool everyone is raving about. We couldn’t agree more!
Keep an eye on our socials and release emails to stay informed. In addition to new features, we will also start working on some new collaborations with the people we met at the conference. We cannot say much just yet, but let’s just say that if you are interested in GPUs and Computer Vision cases, you will definitely like what’s to come.