Give me a Hug(ing Face)

When did you start that?

With so many new community-led projects starting every day its hard to keep up. I wanted to give a quick round up of all the cool stuff we've got going on, in case well you know, you missed it.

Meetups - Live events happen every Wednesday at 5pm UK/9am PST. Join the zoom call here lead on this @Demetrios
Coffee sessions - tend to get released once per week (recommend subscribing to youtube or spotify to stay up to date) lead on this @Vishnu Rachakonda
Reading club - Happens every 2 weeks. More info here and stay up to date in #reading-group lead on this @Charlie You
Engineering Labs - A new cohort is launched every 5 months. We just finished the most recent one. stay up to date in #engineeringlabs lead on this @IlNardo92
Office Hours - We meet every Thursday at 4:30 UK/ 8:30am PST. No recording, no fluff. Just people talking about what they are working on and learnings they can share. stay up to date #officehours lead on this @Harry Berg
MLOps Stacks - Open source project to help make plugging together tools easier. Follow or contribute in #mlops-stacks lead on this @Luke Marsden
Women of MLOps - Feminine energy united! Stay up to date in #women-of-mlops lead on the @Aparna Dhinakaran
Tooling Working Group - A working group that is helping create the frameworks for tools in each part of the ML Dev lifecycle. Lead @Demetrios
AIIA - The AI Infrastructure Alliance. An alliance of ML tooling companies with the goal of establishing a canonical stack for MLOps. Lead @Dan Jeffries

Apart from this, we have started to create channels for different tools and oss projects. If you do not see a channel for a tool you want to know more about or talk about, let me know and I will make one!

Deploy MLflow in an AWS app runner

Community member Douglas Trajano wrote a nice blog tutorial on that very topic! Check out some excerpts below.

'AWS launched in May 2021 the AWS App Runner, a fully managed service that makes it easy for developers to quickly deploy containerized web applications and APIs.

Unlike AWS Fargate, AWS App Runner has a pricing policy based on requests that your application receives. Apps that don’t receive requests all the time can benefit from it.

Our loved MLflow fits perfectly in this context. The interaction with MLflow is usually made in training jobs, data scientists evaluating their experiments, or APIs that expose our models. In all these cases we don’t have frequent access, it occurs eventually.'

The CTO of Hugging Face

David Aponte came out from being held hostage in his mother in laws dry cleaning business and made a cameo on this week's coffee session. Oh yeah and we also talked to the CTO of Hugging Face - every data scientist's favorite NLP Swiss army knife.

Julien, David and I spoke about many topics including:

Infra for hosting models/model hubs
inference widgets for companies with CPUs & GPUs (for companies)
Auto NLP which trains models
"infrastructure as a service"

It was a great talk with none other than Julien Chaumond. Also, David blink 2x if you need us to come get you.

The SAME Project

One of my favorite people to talk to in the MLOps space is David Aronchick. His vision for where the ecosystem is going and what he expects from this space facinate me. It doesn't hurt that he has a pretty solid track record too. (He is one of the people that created Kubeflow)

Well this week he is going to be presenting his latest and greatest endeavor, The SAME Project: A Cloud Native Approach to Reproducible Machine Learning. We all know how important reproducibility is. In fact, I would even go as far to call it a foundational pillar of MLOps.

Join us Wednesday aka tomorrow at 9am PT/ 5 pm BST. In case you want to keep up to date, subscribe to our public google calendar.

Tons of great threads and posts this week! Just a sampling contained here:

Thanks to community member Prashant Rathi, we had a great thread about who in the org really "owns" MLOps. Check out the thread for some great responses!
Shout out to Tecton CEO Mike Del Balso for sharing his commentary on a great Erik Bern blog post! Some very exciting content coming soon...
Great blog post by Kyle Gallatin of Etsy, a previous Coffee Session guest about how to deploy ML models in 5 different ways.
Some MLOps funnies to bright your day... this community can be hilarious!