Seeing into the ML crystal ball

As you are probably aware I absolutely love trying to crack the ML observability puzzle as it is so different from your classic DevOps traces, logs, and metrics. A few weeks ago we had Lina on the meetup to talk to us about some best practices she had found in regards to ML monitoring. This week we build on that by having Aparna Dhinakaran on the meetup to give her take.

This talk will highlight common challenges seen in models deployed in production, including model drift, data quality issues, distribution changes, outliers, and bias. The talk will also cover best practices to address these challenges and where observability and explainability can help identify model issues before they impact the business. Aparna will be sharing a demo of how the Arize AI platform can help companies validate their model's performance, provide real-time performance monitoring and alerts, and automate troubleshooting of slices of model performance with explainability. The talk will cover best practices in ML Observability and how companies can build more transparency and trust around their models.

*IMPORTANTE*
This meetup will take place 1 hour later than normal. Wednesday 6pm UK / 10am PST

On Thursday morning at 10 am UK we will have a bonus meetup with Danny Ma coming to us from Sydney Australia. (you may have heard of him, he is kinda a big deal on Linkedin, and I'm a little nervous to interview someone at influencer status 😅)

Most MLOps discussion traditionally focuses on model deployment, containerization, model serving - but where do the inputs come from and where do the outputs get used?

In this session, we will demystify parts of the data science process used to create the all-important target variable and design machine learning experiments.

We will discuss some probability and statistical concepts that are useful for MLOps professionals. Knowledge of these concepts may assist practitioners working closely with data scientists or those who aspire to build complex experimentation frameworks.

Something really cool happened last week, two community members took the lead on creating a coffee session without me. Wish I coulda been there but I give the guys an A+ on the content of this coffee session. David Aponte and Ryan Dawson sat down for a low key chat centered around why MLOps is not just another flavor of DevOps.

It can be tricky to explain MLOps to colleagues and managers who are used to traditional software engineering and DevOps, let alone your gran. We have to answer the 'Isn't that just DevOps?' question clearly, otherwise the challenges of MLOps will continue to be underestimated (potentially by us as well as others). In this session, they dive into what is new about MLOps and why current mainstream DevOps alone does not solve the problems.

Check out the video recording here and the podcast here