Big events draw in an array of individuals to learn and connect, and the Open Source Summit North America is no exception. Offering a multitude of events and sessions, attendees will be immersed in all things open source.

Taking place at the JW Marriott Austin in Texas, open source enthusiasts will have the chance to collaborate in person after two years of attending the conference virtually. Jam-packed with sessions to uncover from June 21 to 24, be sure to catch three talks by Collabora's very own.

Highlighting the benefits of a people-first driven approach, People Lead Manager Dave Bevan will be giving two separate talks. The first will breakdown the importance bridging cultural distinctions and the second will showcase strategies to foster team spirit within a remote environment.

Senior Software Engineer Jakub Piotr Clapa will speak on how errors can slip through data sets and methods the Machine Learning team have tried to combat this.

Collabora @ Open Source Summit North America

Yes there are differences but we are all human right?
Presented by David Bevan - Tuesday, June 21 – 5:00 pm UTC

How many people are there in the world? The famous book Hitchhikers guide to the galaxy may tell you the answer to life the universe and everything is 42. The answer to the population question though is 166,666,666 times more than that. There are over 7 billion of us on this planet. So why is it so hard to recruit and retain great people? In the talk, Dave will share strategies and initiatives he has employed through workplace culture to hiring and filling the recruitment pipeline to tackle the issue of a diverse workforce.

Remote doesn’t mean distant - managing teams effectively from different geographies
Presented by David Bevan - Wednesday, June 22 – 4:50 pm UTC

Dave Bevan manages 29 people across 13 countries and spanning 7 different time zones. He also only speaks one language; I guess we should wish him luck. Yet within his first 6 months at Collabora, the feedback he has received includes being the best manager these 29 people have ever had. Dave will talk about initiatives he has introduced to help him connect with people remotely and encourage and build a sense of team and belonging within his team of engineers. As a leader within the community, he hopes you care enough to come and join the conversation.

Labeling tools are great, but what about quality checks?
Presented by Jakub Piotr Clapa - Thursday, June 23 – 7:55 pm UTC

Data sets are the backbone of Machine Learning (ML), but some are more critical than others. There is a core set of them that researchers use to evaluate machine-learning models as a way to track how ML capabilities are advancing over time. Over the years, studies have found that these data sets can contain serious flaws. ImageNet, for example, has several labels that are just flat-out wrong. In this talk, we will share our findings on how label errors affect the existing training process, discuss possible implications, and dive into how we leveraged unsupervised learning to filter out annotation errors while looking at real-world examples.

Pop by after the talks for a quick hello, our presenters would be happy to connect. If you can't make it to the in-person events, there are still tickets to attend virtually!

See you next week!

Related Posts

Finding the secret ingredient at Embedded and Kernel Recipes

Talks of the town: Software engineering edition

School's back in session at Open Source 101

Related Posts

Finding the secret ingredient at Embedded and Kernel Recipes

Talks of the town: Software engineering edition

School's back in session at Open Source 101