🏴 MLOps + Tofurky?

Background: I realize that I have the benefit of meeting and hearing about many tools that are coming out in the ML tooling ecosystem weather that is from people reaching out to me, or just cause I spend an absurd amount of time researching this stuff. So I thought I want to create a section of the newsletter for any time I hear about a new tool that I think could help the greater community. These posts are NOT sponsored and you cannot get me to do a feature on you (so please don't ask). The whole point is to shine a light on smaller lesser-known tools that I randomly come across. I will not be doing this in every newsletter, just when something grabs my eye.

The Tool

Replicate.ai caught my attention last week because they are attacking the classic problem of reproducibility within ML. The founding member of this open-source project is Ben Firshman the same guy that created docker compose.... so they got that going for them.

Why tackle this problem, now?

As they put it on their Github page, everyone uses version control for software, but it is much less common in machine learning. Why is this? We spent a year talking to people in the ML community and this is what we found out:

Git doesn’t work well with machine learning. It can’t handle large files, it can’t handle key/value metadata like metrics, and it can’t commit automatically in your training script. There are some solutions for this, but they feel like band-aids.
It should be open source. There are a number of proprietary solutions, but something so foundational needs to be built by and for the ML community.
It needs to be small, easy to use, and extensible. We found people struggling to integrate with “AI Platforms”. We want to make a tool that does one thing well and can be combined with other tools to produce the system you need.

I am anxious to see if they can set themselves apart from other tools out there that already exist, mainly DVC. I know they aren't the only ones who have tried to create a culture of version control in the ML world, so I wonder if it will stick this time. When I talked to Ben he told me "the primary thing we are trying to do right now is create community because we think this thing should be built by a community not, just a single vendor".

The How: Nathan and Tim were absolutely incredible to talk to about how they see the current MLOps landscape and what their vision is for the future. I thoroughly enjoyed getting to ping pong the questions off both of them and hear what specific takes they had on what tools on the market were getting them excited these days.

Favorite quotes from each of them:

Nathan making the case for buying tooling - "Fast forward a few years when your company has multiple geographies, multiple product pipelines, customer archetypes - Suddenly your internally built MLOps tool is no longer a priority. Freeing up some developer time which can be replaced by a SaaS that has its own startup behind it is better value for the money."
Tim talking about how the MLOps tools have changed recently - "MLOps is moving along with the ML Maturity of companies using Machine Learning products. The further along these companies are they start to think about more issues and have more needs."

More Good News: Both these cool cats are in our community slack so feel free to reach out to them and start a conversation. Check out the full conversation on youtube or in the podcast lands

For the third installment of our data privacy series, we touch on ML topics that go far beyond the technical mumbo jumbo. We all know the potential of AI and the amazing transformations it can bring, but are we prepared for the changes? Do we really understand what bias and fairness stand for when it comes to AI? Can we enforce AI compliance?

The Ethics behind AI is a new trending topic with many hard questions still left unanswered.

Special Guest: For this episode, Fabiana talks with none other than Mr. Charles Radclyffe himself in the flesh! If that name sounds familiar it's probably because he was our 2nd ever meetup guest way back in April when this whole thing kicked off. Charles is a serial entrepreneur and a data philosopher who has focused his career on solving tough tech challenges for some of the world’s largest organizations.

This series is brought to you by YData. YData offers a dataset experimentation platform with synthetic data generation that makes the process of building datasets take a fraction of the time and cost that they used to.

The Lowdown: David got his nose broken back into the right position so now A. he doesn't look as funny B. he can breathe better - but you know what that means? He wasn't around for our last coffee session. And I was feeling lonely.

Not to fear, I didn't go in all on my own, I called for reinforcements and community member Vishnu Rachakonda bravely stepped forward! All jokes aside though, I think Vishnu might have to join us for a few more cause he was such an awesome host!

But What Did We Talk About?
Vishnu and I spoke with Barr Moses the CEO of Monte Carlo Data about one of my favorite topics; Monitoring! This time more specifically, data monitoring!

We touched on:

The rise (and threat) of data downtime
The relationship between DevOps Observability and Data Observability
Data Observability and it's five key pillars
How the best data teams are leveraging Data Observability to prevent broken pipelines

Check out our full conversation on youtube or in podcast land and give a big shout out to Vishnu next time you see him!

Backstory: In August, we announced Engineering Labs. During these last couple months we have been thinking hard about how to structure this initiative in the right way. Now we are ready to open up. Join us for the big reveal next Wednesday, December 2nd at 5pm UK/9am PST.

What Is It?
We are going to talk about Working guidelines for our first voyage with you all! I know many have been asking for hands on experience with MLOps so this is a perfect excuse to test your abilities! jump into the zoom call and come see what all the fuss is about!