Just Build It! Tips for Making ML Engineering and MLOps Real

Data science and Machine Learning in an industrial setting are hard. The problems you have to solve are complex, the data landscape is challenging and you often don't have the freedom you would like to design experiments or create observational studies on real-world processes. This is before you even think about how to manage stakeholders, use cloud technologies, write software or wrap your solution up into a product that has to run predictions 24/7/365 and support business operations! In this talk, we reflect on many of the learnings Andy has gained through the past few years working in successful data science and machine learning engineering teams building operational products that create millions of dollars of value. In particular, Andy discusses how he thinks we can 'bootstrap' ML Engineering (MLEng) and MLOps practices in your organization.

Take-aways

In detail Andy will cover topics like: - Where do you start? How do you know what to go after first when it comes to operationalizing your ML models? - What comes first - process or tooling? If you're trying to make ML Engineering and MLOps real, do you need to work out good processes first or do you need to invest in the right tools first? - Who do I need? If you're building out an MLEng/MLOps team, what mix of skills and capabilities should I have? How should I organize them? - How do we do it? On the ground, what do we need to bear in mind as we develop our solutions? How do we set ourselves up for success? Some of the points covered are also covered in Andy's book, Machine Learning Engineering with Python (https://www.amazon.co.uk/Machine-Learning-Engineering-Python-production/dp/1801079250). You should come away from this talk: - Feeling confident in some of the key questions you'll need to tackle to get MLOps 'actually working'. - Knowing the main tooling landscape and processes you could employ to get MLOps up and running. - With some feeling of some of the pitfalls to avoid! - With pointers to further detailed technical material and resources to back up the content shared in the talk.

In this episode

Andy McMahon

Machine Learning Engineering Lead , NatWest Group

Andy is a machine learning engineer and data scientist with experience of working in, and leading, successful analytics and software teams. His expertise centers on building production-grade ML systems that can deliver value at scale. Andy is currently an ML Engineering Lead at NatWest Group and was previously Analytics Team Lead at Aggreko. He has an undergraduate degree in theoretical physics from the University of Glasgow, as well as Masters and Ph.D. degrees in condensed matter physics from Imperial College London. In 2019, Andy was named Data Scientist of the Year at the International Data Science Awards. He currently co-hosts the AI Right podcast, discussing hot topics in AI with other members of the Scottish tech scene.

Twitter

Demetrios Brinkmann

Host

Demetrios is one of the main organizers of the MLOps community and currently resides in a small town outside Frankfurt, Germany. He is an avid traveller who taught English as a second language to see the world and learn about new cultures. Demetrios fell into the Machine Learning Operations world, and since, has interviewed the leading names around MLOps, Data Science, and ML. Since diving into the nitty-gritty of Machine Learning Operations he felt a strong calling to explore the ethical issues surrounding ML. When he is not conducting interviews you can find him making stone stacking with his daughter in the woods or playing the ukulele by the campfire.

Take-aways

In this episode

Andy McMahon

Demetrios Brinkmann

Upcoming episodes

Declarative MLOps - Streamlining Model Serving on Kubernetes

Rahul Parundekar

Tecton 0.6: Notebook-driven Development

Jason Dunne