Jamf’s first internal Data Science conference!

David Pryce-Compson
Jamf Engineering
Published in
5 min readAug 16, 2022

--

Jamf’s Data Science team was born with the acquisition of Wandera just over a year ago. Our team is primarily based in EMEIA, but with a growing presence in North America. The main focus of our work are Jamf’s security products, but we are getting more involved with the products in the Jamf management portfolio too.

DataCon Brno 2022 was born as a way to bring our team together, to get to know each other better over a coffee or a beer and to invest in our growing Data Science practise. Here is a summary of the why and the what of DataCon Brno 2022, and also how we plan to grow it in the future!

What DataCon looked like

We knew we had to get as many of the team in one place, so it made sense to base DataCon in Brno, home to the majority of the team already. We also knew we wanted some structure for the event, so planned in the following

  • A hackathon — to allow the team to innovate in a focused environment
  • Knowledge sharing — to allow data users at Jamf to show off some of their recent and upcoming work to the rest of the company
  • Team building — intersperse with team lunches, coffees, drinks and a one afternoon team building activity (we did an escape room and wine tasting)
  • Merch —the most important part, right?
DataCon logo and MIRIAM stickers with coin for scale
All good conferences need merch, or at least stickers …
… and tee shirts!

The Hackathon

We set aside two days during our visit for the hackathon, preselected some ideas and we made sure to appoint an unbiased judge to decide a winner.

The two ideas we went with were selected on the criteria of being either something that would make our job and hence the cycle of research to development easier OR a feature that could be used by our customers straightaway. These were:

  • URL analysis library — the core data for most of our Data Science projects (stay tuned for a blog post very soon!)
  • Using HTML to analyse our data — a new idea that could be used in conjunction with the first one to help complement our existing detection or classification models (stay tuned for a blog post very soon!)

We divided the team into two groups, so that each would develop one of the ideas. At the end of the two days, both groups presented their work to our Cloud Engineering Director, Enrique Cornago, who added valuable feedback into the potential applications of both these mini projects.

The winning team was the URL analysis library for its ability to truly be extensible and used throughout the Data Science initiatives and even begin to extend our content filtering capabilities across the world i.e. in various languages. They won a round of beers from the losing team!

Two male Data Scientists hacking away at laptops, sitting at a table next to a white board, one facing the camera, another hiding under his green baseball cap
HTML profiling team
Two Data Scientists, one female, one male. Facing away from the camera. Sitting a long table, hacking on laptops. Next to a whiteboard.
URL python library team (winners)

The Conference

We always focus our time on doing the job at hand but often forget that others in our organisation want (and even need) to learn about what we do and how it works, even if only at a high level.

The features and data that drive decisions within and around our products can help Developers as they create them, Marketing as they wax lyrical about them and Sales of all types as they demo and sell them. Hence having an expert educate with clarity is key to the success of these features and initiatives. DataCon gave us the perfect opportunity for this.

Not only did the Data Science team discuss how we work and some of the awesome things we have done lately or plan to do soon; the ThreatLabs and DBA teams were invited to present on their fascinating contributions also (big shout out to Michal Rajcan and Allan Briffa).

From start to finish we had a lot of interest in these talks, with on average 30, max 60, people attending each talk either on Zoom or in person. Overall it was a very rewarding experience.

A white male in a burgundy tee shirt and shorts, standing and presenting in front of a projector and laptop showing some technical information
Michal Rajcan presents on how ThreatLabs uses data.

Lessons learnt, 2023?

Of course we can look back at it now and if given the chance would do something differently or to a larger scale:

  • Bring in as many of the people who might want to contribute either in a hackathon, ideation or presenting as early as you can
  • Offer multiple tracks for your conference to ensure everyone else can pick and choose their topics
  • Recordings and documenting like this blog post are super important to include everyone even if they couldn’t make the actual event due to work or timezone
  • Hackathons might be better with two days plus an added half day to prepare / wrap up
  • Always make tee shirts and fun stickers — it’s fun, commemorative and good internal marketing (Ed: always recruiting aren’t you!)
  • Try to get the whole team in the same place
  • Time together as a team is very enjoyable!
Six people smiling at the camera in matching white tee shirts with Data Con logos on them. There is a large team photo in the background.
Can you tell who MI:RIAM had to photoshop in / who missed the photo op?
Four adults sitting at a table smiling towards the camera with glasses and candles in the foreground
Team dinner / wine tasting
An underground wine cellar, well lit, with tables and long seats on both sides.
The wine cellar

At heart this was about bringing an innovative and like-minded team together but it also allowed us to show the importance of coming together and seeking out those valuable next steps and who knows, this might be the start of a new internal annual conference. Maybe somewhere more tropical next time?

--

--

David Pryce-Compson
Jamf Engineering

Mathematician / Data Scientist, Ultimate player / coach / commentator, occasional climber