Welcome New Friend

Vishnu couldn't make it this week but Demetrios managed to convince Skylar Payne to come co-host with him. It was especially fitting because our guest of honor was Skylar's mentor at LinkedIn when they worked together.

Alex Pantry walked us through the transformation LinkedIn underwent over the past two years on their job recommender system. This talk hit hard, Alex brought his 'A game'. Skylar could dive in deep from his experience working with Alex.

Patry has several years of experience leading large groups of machine learning engineers at LinkedIn. In his time at LinkedIn, Alex has mentored many engineers, solved problems of scaling machine learning across many teams, and made bold bets on how to take LinkedIn's search and recommendation to the next level.

Alex shared his wisdom on making bold bets: from de-risking the investment to gaining stakeholder alignment as well as a lot of insight on what it means to be a good machine learning engineer. Alex also drops a ton of fortune-cookie worthy lines in this chat, so come for the knowledge and stay for the cookies

Anecdotal data also tells a story
AI engineers are like water, they go where there's no friction
Innovation happens when you empower people
Every good MLE should be able to tell a story with the data

End to End MLOps

Raviraja is currently working at Enterpret as a Founding Engineer - NLP.

In this talk, he takes us through the high level steps in the MLOps cycle, provides an example tool for each step and how it all fits together. He put together a learning path for himself to learn about all the pieces of MLOps and then made sure to document it so its available to all of us.

Check out a detailed example by going to his Github where you can find a link to a blog post for each step and the code to get it up and running. You can also access the slides he used for the presentation.

Alex Serban, PhD Candidate from Radboud University, works at the intersection of machine learning and software engineering, looking for ways to design, develop and maintain robust machine learning solutions.

After talking with over 500 of practitioners, he compiled a catalog of engineering best practices for the development of ML applications. For this talk he gave an overview of the findings and presented a quantitative method to assess a team’s engineering ability to develop software with ML components and suggest improvements for your team’s processes.

Kserve

In light of the new graduation Kserve veteran and second time meetup guest, Theo Papapanagiotou is going to go deep on what the hell is going on with the artist formerly known as KFserving.

What to expect?
We will start with the serialization of TensorFlow/PyTorch/SKLearn models into files and the deployment of an inference service on a Kubernetes cluster. Great MLOps means great model monitoring, so then we will look at inference service metrics, model server metrics, payload logs, class distributions.

For AI ethics in production, we will use the explainers pattern with many different explainers, fairness detectors, and adversarial attacks. For integrations, we will use the transformer pattern to process as well as to enrich the inference request with online features from a feature store. Finally, we’ll look at how to build a custom inference service using the KServe SDK.

By the way.... have you heard the MLOps Community has a public calendar? Subscribe to it or click the button below to jump into the meetup on Wednesday at 9am PST/5pm BST