Join us for a night of Python talks at the central Skyscanner office in Glasgow. We will have two talks (details below). Skyscanner will be providing us with food and drinks.
Doors open at 17:30, please arrive by 18:30 at the latest.
Pizza will arrive at 18:00.
First talks will start at 18:30.
Please note: 155 St Vincent Street reception is locked from 17:30 onwards. To access the building please wait patiently by the front door a member of Skyscanner will then be able to unlock the door and provide access. Access can only be guaranteed up until 18:30.
Talk 1: APIs On The Scale of Decades by Gary Fleming
"APIs are hard. They are pretty much ship now, regret later." -- Chet Haase.
What do Greek philosophy, early video games, and Japanese bullet trains tell us about how we should design our APIs?
Writing any old API is easy. Writing an API that can evolve to meet your needs over the coming months, years, and even decades; now that's hard. We'll look at some common practices and try to see where they go wrong, some misunderstood techniques and how to use them better, and some less common practices that might be useful.
Let me give you some good advice that'll help you evolve your APIs, and some big ideas that might provoke some interesting thoughts.
Join us for a night of Python talks at the central Skyscanner office in Glasgow. We will have two talks (details below). Skyscanner will be providing us with food and drinks.
Doors open at 17:30, please arrive by 18:30 at the latest.
Pizza will arrive at 18:00.
First talks will start at 18:30.
Please note: 155 St Vincent Street reception is locked from 17:30 onwards. To access the building please wait patiently by the front door a member of Skyscanner will then be able to unlock the door and provide access. Access can only be guaranteed up until 18:30.
Talk 1: APIs On The Scale of Decades by Gary Fleming
"APIs are hard. They are pretty much ship now, regret later." -- Chet Haase.
What do Greek philosophy, early video games, and Japanese bullet trains tell us about how we should design our APIs?
Writing any old API is easy. Writing an API that can evolve to meet your needs over the coming months, years, and even decades; now that's hard. We'll look at some common practices and try to see where they go wrong, some misunderstood techniques and how to use them better, and some less common practices that might be useful.
Let me give you some good advice that'll help you evolve your APIs, and some big ideas that might provoke some interesting thoughts.
Talk 2: Machine Learning in Production with Python-Based Microservices by Andy McMahon
Andy is Head of Data Science & Machine Learning at Streamba, a start-up in Glasgow which develops cutting edge software solutions for the energy industry. His work in the VOR team spans the spectrum of predictive, prescriptive and descriptive analytics with a particular focus on production ready machine learning solutions. He's a keen Pythonista and does almost all of his development in Python. Recently he has been exploring Scala and Java for developing machine learning system.
Description:
Among the many challenges I've faced at Streamba is how do you take your shiny new ML model and actually put it into production? In this talk I'll discuss some of the different ways you can do this, with a focus on the cloud-based microservice architecture. I'll then (time permitting) show a running example of this with a very simple ML microservice running in the cloud. Hopefully you'll come away with a better idea of how to take your Python scripts and ML code and convert them into a running service that brings value to customers and internal stakeholders.
