Hands Down The Easiest Way To Do MLOps

Billions Of Perdictions Per Day

Not only was Manoj all smiles on his birthday but he gave us some amazing details about what the salesforce ML platform Einstein has achieved thus far and how! I couldn't help but ask if his team was just going to parade all these feats around to the internet and make the rest of us jealous or if it would ever be available to the greater public? Spoiler alert: they are going to open source as much as possible!

Favorites
Honestly, this talk was pure gold. Manoj took us through the requirements that his team needed to accomplish with respect to model serving, feature engineering, and model training.

As if that wasn't enough, he then walked us through some of the problems they encountered after getting everything running, and how they were able to overcome the unforeseen hurdles.

Shuffle Shard Dance
Shout out to the whole Salesforce ML infra team here for the incredible name of how they overcame the noisy neighbor's problem. Honestly, if someone can make the Shuffle Sharp Dance the next best thing on TikTok, I can die with a smile on my face!

Check out our full conversation on youtube or podcast. Due to the nature of this talk (slides and visuals) I, would highly recommend you watch on youtube.

Chow For Now

After a season of 9 episodes all about privacy-enhancing technologies, we bring you the last episode.

Synthetic data can be used as a privacy tool for Data Science teams. Since this is the last episode we've decided that it should be a special one, for that reason, we've invited not one but two of the top minds in synthetic data research - Jean-François Rajotte, Researcher at the University of British Columbia and Sumit Mukherjee, Senior Applied Scientist at Microsoft's AI for Good lab.

A chat about how synthetic data can take healthcare to the next level of AI adoption. That's all though, we also touch on how it can be combined with Federated Learning and Differential Privacy to boost your data privacy preservation.

Sound good? have a listen and also check out some of Jean and Sumit's latest work.

A “Gift” From Above
This week, Demetrios and I got to spend time with the inimitable Noah Gift. Noah is a data science educator, who teaches at Duke, Northwestern, and many other universities, as well as a technical leader through his company Pragmatic AI Labs and past companies.
His bio alone would take up this section of the newsletter, so we invite you to check it out here, as well as the rest of his educational content. Read on for some of our takeaways.

HOW Is As Important As WHAT
In our conversation, Noah eloquently pointed out the numerous challenges of bringing ML into production, and especially for making sure it's used positively. It’s not enough to train great models; it’s important to make sure they impact the world positively as their productionized. How models are used is as important as what the model is.
Noah specifically commented on externalities and how’s it incumbent on all MLOps practitioners to understand the externalities created by their models.

Just Get Certified
As an educator, Noah has seen front and center how deficits in ML/DS education at the university level have led to the “cowboy” data scientist that doesn’t fit into an effective technical organizational structure. In his courses, Noah emphasizes getting started with off the shelf models and understanding how existing software systems are engineered before committing to building ML systems.
Furthermore, Noah suggested getting certifications as a useful way of upskilling for anyone looking to increase their knowledge base in MLOps, especially by cloud providers.

Tech Stack Risk
Finally, as many of you do, we debated the relative merits of the major cloud providers (AWS, Azure, and GCP) with Noah. With his vast experience, Noah made a great point about how adopting extremely new tools can sometimes go wrong. In the past, Noah adopted Erlang as a language used in the development of a product. However, as the language never quite took off (in his experience), it became a struggle to hire the right talent to get things done.

So here is a question for all of you, as you go about designing and building the MLOps stack, does any part of the process sound like Noah’s experience with Erlang? Tools or frameworks where downstream adoption may end up fractured? We’d love to hear more!

Check out the whole convo on youtube and podcast land.

Yo Adrian!

Machine Learning Design Patterns has become an instant classic for those of us interested in MLOps. After speaking with Sara Robbinson one of the authors we now get to talk with another author but this time in front of a live audience aka you!

More Patterns
In our first convo about this book we talked mostly about the pipeline pattern and in this iteration of the conversation I'd like to dive into Model versioning, responsible AI, and the hottest topic of 2020; feature stores. I also would like to touch on the idea around ML maturity, and how necessary is automation when dealing with ML, especially during the research phase.

+ I am leaving the agenda a bit open because I know there will be a few of you that also have questions for Lak. Come prepared!

+ Some of you have been asking me how to get the meetup added to your calendar. If you use google calendar you can add it here and outlook here. Let me know if you have any problems.

+ See you at 5pm GMT / 9am PST tomorrow Wedensday by clicking the link below.

Nvidia Jetson Nano Developer Kit B01

Community members Vesi and Marian reached out to me to let me know they are raffling off an Nvidia Jetson Nano Developer Kit B01 so I Naturally said we should offer it to you all in the community! So here are the details Vesi sent me over:

We're looking to speak with experts who have succeeded to operationalize at least 1 ML model in their experience. We're conducting short 30-minute online coffee sessions with each person, asking questions about their experience, the challenges they have faced in the process, and how they managed them. The only 2 requirements to qualify are:

Practical experience with ML deployment.
Good speaking level in English.

On February 7th, we'll draw one person's name and reward him/her/nonbinary with the prize! To get in touch with Vasi ping her directly on slack @vesi staneva or hit the button below to send her an email. Best of luck!