Engineering MLOps

Product Vision
We started out strong and kept up the incredible pace last week. Right from the get go I knew this was going to be an incredible session. Korri came out the gates with sound advice answering the question about how we can effectively create products that fill our needs in the present while keeping an eye on the future.

So many takeaways from these product experts. I loved the camaraderie we had. One of my favorite questions was how PM is different when designing for an internal customer as opposed to something you sell on the market.

Here is a link to the video and the podcast.

Privacy, Mistakes and a bit of Synthetic Data

I was going back through the limited series we did with Fabianna Clemete on Data Privacy and wanted to write a blog about some key takeaways. As most of you know I also do a weekly podcast on the ethics of ML/AI so this blog was right at the crossroads of my interests.

On another note, as I write this I know synthetic data isn't all that common. I would love to hear from you about synthetic data. Is it fixing a need? What are your thoughts? Is it a valid solution to a very real problem? Or....

Tribe AI Consulting without the Hustle

Tribe AI is a curated community of ML engineers, researchers, and data scientists that partner with companies across industries on AI projects that drive business impact. Tribe is made up of entrepreneurs, nomads, parents, and volunteers who have ditched the corporate 9-5 to work on impactful AI projects on our own schedules.

We talked to Oleksandr Paraska, a member of Tribe and MLOps community, about why he wants to spread the word about Tribe.

Why do you think Tribe is a great fit for someone in the MLOps community?
Tribe lets you dial up and down your consulting depending on what else you have going on. I know there are a lot of people in ML Ops like me, who are starting their own company or interested in going that route. I think 90% of the people I know would benefit from consulting on the side and then there’s the community aspect of Tribe too.

What’s the Tribe community like?
Once I got into Tribe, I was immediately impressed by the quality of talent. There are a lot of communities out there, a lot of Slack channels with a thousand people. Tribe isn’t that. You have to apply and everyone is vetted, so it really feels like a community you want to be part of. We’re all remote, but we get to work together on projects, which is fun.

What’s been most helpful about being part of Tribe?
Tribe can provide different things depending on what you’re looking for. Consulting is one of them, but also connections to these talented people and even investment opportunities. Even without the consulting work, for me, the community would be a draw. It’s all just emerging, which is what makes it exciting, but I love what we’re building.

Has it been helpful for your startup?
My company builtup.ai is helping orgs take their first steps in AI. After my experience at the Founder Institute’s accelerator, I realized I needed to get inside more organizations and see the challenges they faced trying to roll out AI. Consulting with Tribe has been a great way to talk to more users and get more insights.

What’s different about working with Tribe versus going it solo?
Tribe takes all the sales work off your plate – they do almost everything. And they’re great at presenting you in the best light to clients and helping you get the rates you want.

Tribe AI is fully remote and accepting applications from ML engineers, researchers, data cowboys, and product managers from around the world. If you want to do cool ML projects, make money, and work the way you want – get in touch.

Sound Familiar?

You may have seen this book already. Emmanual Raj the author was kind enough to give away 5 copies of the book to the community. Or maybe you have seen it making waves in the MLOps space. Whatever the case, I convinced Emmanuel to come on the meetup this week and talk to us.

The talk will focus on simplifying/demystifying MLOps. It will encourage others to take steps to learn this powerful SE method. We will talk about his journey into ML engineering, the evolution of MLOps, daily life, and SE problems. To top it off we will dive into what's next in MLOps and since many of you asked we will also do some live coding! Whooohoooo!

I know we haven't had the talk yet but this is what I am aiming to uncover in our chat:

Why ML projects fail (based on Emmanual's experience)
Some of the principal challenges in productionizing ML
How is MLOps a saviour (projects succeeding)
What is AIOps (automated error handling in DevOps)
MLOps process simplified
MLOps tools recommendations
Explainable Monitoring (production monitoring framework and techniques)
Next steps for killer ML solutions- AIOps + MLOps

See you on Wednesday aka tomorrow at 5pm UK / 9am California. Click the button below to jump into the event, or subscribe to our public google calendar.

Laszlo Strikes Again: It's a recurring discussion--do data professionals need to write high quality code? Laszlo Sranger provides a compelling explanation for why.
Managing Data Science Teams: A great thread in the #production-code channel highlights how to manage across the objectives of product, data, and engineering and set teams up the right way.
Metrics for Data Engineering: A really useful set of metrics for measuring the impact of data engineer teams.
Cool event from Anyscale: Loving the lineup of strong companies and practitioners at the Ray Summit. Presenters from Robinhood, Shopify, Uber, and more are sure to present great content.