Past Meetup

Productionising Data Science

This Meetup is past

198 people went


Join us for our next ASI Meetup, located at the Royal Statistical Society on putting Data Science into production.


6.00pm – 6.30pm: Registrations with Drinks/networking

6.30pm – 7:00pm: Dr. Pascal Bugnion & Dr. Scott Stevenson (incl Q&A) - ASI Data Engineering Team

7.00pm - 7.15pm: Lightning Talk: Dr. Victor Zabalza (incl Q&A) – ASI Data Engineering Team

7.20pm - 8:00pm: Finish and drinks/ Pizza & networking


Sign up

***Please note that we do need the name of everyone coming to the meetup to provide for the security at the venue for you to get in, to do so, please register via this link Eventbrite for your ticket. We also have limited space so please update your eventbrite availability to allow people to go in your place if you can no longer make it. Those without an eventbrite ticket may be turned away.


Biography: Pascal Bugnion

Pascal is a Data Engineer with an analytical background as well as a passion for software development. He pursued a Materials Science undergraduate at Oxford University, and awarded his PhD in computational physics at Cambridge University, during which he developed quantum Monte Carlo methods to solid-state physics. He has authored many publications, and also recently released a textbook on the programming language Scala. Pascal has also contributed to NumPy, Matplotlib and IPython, and maintains scikit-monaco, a Python library for Monte Carlo integration.

Biography: Scott Stevenson

Scott is a Data Engineer with strong software development skills who contributes to a number of open source projects. He has worked on a range of complex problems, including his ASI Fellowship project designing and building a scalable engine for scheduling and running custom data loading and processing jobs. He completed his DPhil in Particle Physics at Oxford, carrying out statistical analysis of multi-terabyte datasets collected with the Large Hadron Collider at CERN.

Talk Synopsis: From Jupyter notebooks to live systems

You have just written a Jupyter notebook that classifies your company's customers into cat and dog-lovers. All your hard work is useless unless you integrate that notebook into your live systems. But how do you handle retraining your model regularly? How do you hook into your company's login page so your model can automatically assign pet preferences to new customers? We explore best practices in productionising research-stage models.


Biography: Victor Zabalza

Víctor is a data scientist with a background in high-energy astrophysics, with ten years of research experience. During his PhD and postdocs in Barcelona, Heidelberg, and the UK, he has studied the origin of gamma-ray emission in systems within our Galaxy, with an emphasis on using both numerical modelling and data analysis of X-ray and gamma-ray observations to achieve maximum physical inference. Victor has been programming in Python for over eight years, and has lead international workshops on how to use the scientific stack and astronomy software in Python.