Ryan Dawson on Open Source Tools and MLOps — A Noonie Nom Interview

Ryan Dawson is a 3x Noonie Nominee and is a top Hacker Noon contributor in the Software Development story category. In this interview, Ryan shares what he's learned about the open source value chain, MLOps, and problem solving with tech vs. people, or ideas.

Have your say!

1. Which 2020 Noonie/s have you been nominated for?

1. Hacker Noon Contributor of the Year - DEVOPS
2. Hacker Noon Contributor of the Year - ENGINEERING Software Testing Writer of the Year
3. Hacker Noon Contributor of the Year - PROJECT MANAGEMENT
   

2. Tell us a bit about yourself.

I've worked across different areas, and it's made me very interested in how different parts of the software world are connected. It's a world of specialists and we rightly celebrate specialist skills, but we also have to use specialist skills in the context of a bigger picture.

My first projects were on small teams but for big companies, so we had all the process and security challenges of big companies and I was working on a lot of it myself.

Then, when I worked on bigger projects, I saw how some difficult barriers arose in workflow, and that made me interested in how everyone could work better together.

I find it especially interesting that some of the challenges can be solved using technology (like DevOps automation) and some are more about people and ideas (like Agile) and others are both.

I started as a consultant at big banks, and then I moved into open source software at Alfresco.

It can be great to have developers as your users, as it can make it easy to share a common understanding. You also feel quite connected, as you're building on top of open source tools, to provide other open source tools, so you get a sense of a whole value chain.

Now I do open source tools for machine learning at Seldon. It's a really exciting field as so much is new and it's moving so fast.

3. Tell us about the things you make / write / manage / build.

I work in MLOps. It's all about making the machine learning build-deploy-monitor lifecycle as smooth as possible. This is surprisingly challenging, as machine learning uses data in very different ways from most software. You extract patterns from data and reapply those patterns to make predictions. So, if the new data you're applying the patterns to is different, or it changes, then you can get really bad predictions.

Think, for example, of data that changes with the season - you don’t want to be recommending coats to your customers when they want to buy t-shirts.

That's one thing we're working on at Seldon - making it easier to monitor for when the data distribution changes and your model might not be applicable to it anymore.

Machine learning models also get built and packaged differently from most software. Our project Seldon Core is all about making it as easy as possible to add a web API around a machine learning model and run it in Kubernetes like any other app. We add in integrations specific to machine learning where we can, like ways of monitoring for data applicability.

4. What are you most excited about right now?

How open source has grown in influence and is now such a big part of the mainstream. It offers a lot of potential for greater industry-wide collaboration and standardisation in software. I think the way that Kubernetes has grown and come to dominate in the cloud platform space is a powerful illustration of this. There was demand in the industry for flexibility and control over the platform and the right mix of actors came together to make it happen.

5. What are you worried about right now?

Climate Change. It's a worry I have in the back of my mind and it comes and goes. I especially worry about how societies could respond to problems like water shortages and crop failures. It needs people to pull together, and these things have the potential to drive people apart.

6. What's the most useful advice you've ever given somebody?

To use the github shortcut keys. I use 't' to find a file by name in the browser so often. I'm surprised how few people seem to know about them, they seem to be a well-kept secret.

7. How has the pandemic changed your life and/or career

My wife and I move our dinner table across the room each day. We call this 'commuting'. I feel I've been pretty lucky as I've been able to move to working from home quite easily. I've even had more time to reflect and think more about stuff I'd like to write about. It has been much harder for some of my family members so I'm aware of how lucky I've been.

8. If we gave you $10 million to invest in one thing right now, where would you put it?

Emitwise, if they wanted the money. I'm an organiser of Tech Ethics London and Mauro Cozzi told us all about Emitwise. Helping companies better track and improve their overall carbon footprint is such a great concept and I think ethics is increasingly becoming a sound business investment.

9. What's an opinion you have that most people don't agree with?

We should embrace conflict more. I'm a big fan of Patrick Lencioni's book 'The Five Dysfunctions of a Team'. Conflict is often taken to be a sign of problems but an absence of conflict can be a bigger warning sign. People need to speak openly and express their thoughts in order for a team to unite around a vision.

10. Which apps can't you live without?

I don't actually use much on my phone. Whatsapp, I guess. On my computer I love command-line tab-completion.

11. What are you currently learning?

In my day job I've been learning more about OpenShift. On the side I've been reading Mik Kersten's book 'Project to Product' and recently read 'The Technology Fallacy'. I'm interested in how software delivery can best realise business value. I'm also doing more presenting - that's a skill I'm trying to get better at.

are Back for round 2, and — in keeping with 2020's general theme - this year's are bigger, badder, & longer than ever:

🚀 Refactored Next.js platform

🚀 $10K+ & 1k+ Domains in Prizes

🚀 2K+ Nominees

🚀