Lytx: Using Data to Help Save Lives on Our Roadways

(photo: Lytx)

Lytx® — as in analytics — harnesses the power of video to transform transportation fleets with improved safety, efficiency, productivity and profitability. Lytx pioneered the video telematics category two decades ago when it launched its DriveCam® safety program, and remains the market leader today. With more than 500,000 subscriptions, Lytx protects over 3,000 commercial and government fleet clients and 850,000 drivers. The company has analyzed more than 80 billion miles of driving data — adding 1 billion miles of new data every two weeks — to deliver unmatched insights into improving driver performance and roadway safety.

“Our flagship DriveCam safety program is considered the industry standard-bearer of driver safety,” said Stephen Krotosky, who leads the Applied Machine Learning Group at Lytx. “The solutions we develop are deeply connected to our core safety mission, and we are using and building our machine learning and AI infrastructure to enable us to automatically identify situations that indicate something unsafe is occurring, or may be at risk of occurring.”

“Labelbox’s infrastructure work enabled us to focus on modeling, design and evaluation of our game-changing machine learning initiatives.”
Stephen Krotosky, Applied Machine Learning Group at Lytx. (photo: Lytx)

“Our team continued to expand as we advanced from labeling a few thousand images across a single project, to dozens of units collecting over 10,000 events per day. We knew we had to scale up the review team to handle the amount of data coming in, and we needed to do it quickly. That’s when we started to work with Labelbox, and they’ve been a good fit for us.”

Labelbox went beyond solving Lytx’s immediate challenges: “Labelbox brought us additional features to help expedite our work. With Labelbox, we can customize a template and quickly get to work.”

“We also use Labelbox to monitor our deployed models in production. We integrate labeling tasks done in Labelbox with our model predictions to perform QA and feed key metrics into our performance dashboard. We A/B test different models and check their performance.”

“We’re pleased with how Labelbox can scale with the growth of our team or expansion of our machine learning initiatives.”
Lytx DriveCam safety program event recorder (photo: Lytx)

“Labelbox has supported us in so many ways, from designing custom interfaces and our user web portal to the complex task of managing multiple reviewers. Their infrastructure work enabled us to focus on modeling, design and evaluation of our game-changing machine learning initiatives.”

“Labelbox is very well integrated as a part of our overall system design. We use it exclusively for research data collection tasks and are pleased with how Labelbox can scale with the growth of our team or expansion of our machine learning initiatives.”

“Labelbox has helped us efficiently collect and analyze data through their strong infrastructure support. We’re able to get the labels we need for all different types of video and image analysis.”

“Labelbox is great to work with.”

Get Started with Labelbox

Visit www.labelbox.com to evaluate Labelbox for free or speak to one of our team members for an enterprise offering.