Lyft Perception Challenge

Brought to you by

enter the challenge

Earn an interview with Level 5, Lyft's self-driving division!

About the Challenge

Lyft is rapidly growing Level 5, their self-driving division. They’re looking for Talented Software Engineers, and they want to talk to you!

Lyft and Udacity are partnering to offer a challenge for engineers who want to work on self-driving cars. To succeed, you’ll need to demonstrate your ability to design a perception algorithm for object recognition and image segmentation. The top 25 competitors will secure an interview with a Lyft hiring representative!

To be eligible for this challenge, you must be enrolled in or graduated from the Udacity Self-Driving Car Engineer Nanodegree program. To earn an interview, you must have U.S. work authorization. The challenge will close June 1, 2018. For submission details and other FAQs see details below.

Winners will be subject to terms and Official Rules.

Our Partner

Lyft is a Leader in the Self-Driving Car Industry

Lyft's Level 5 Engineering Center in Palo Alto, California is a brand new facility focused on building out their self-driving technology and becoming the autonomous platform of the future. Lyft is committing its vast transportation expertise to building safer, cleaner and more efficient communities through the development of self-driving technologies, and this is tremendous news for everyone who shares their vision for a better future.

About Lyft


  • How do I know if I’m eligible to compete in the challenge?
    To be eligible to enter the Lyft Perception Challenge, the minimum requirement is that you are enrolled in Term 1 of the Udacity Self-Driving Car Engineer Nanodegree Program. Students in Terms 2 or 3 are also eligible, as are graduates of the program.
  • I’m eligible! How do I enter the challenge?
    Click the “Enter the Challenge” button to enter. You’ll be taken to the classroom, where you can find more detailed information on the challenge itself and the submission process.
  • I clicked the “Enter the Challenge” button but it just took me to sign-in page. What went wrong?
    You have to be signed in to your Udacity account to enter the challenge. Sign in to Udacity through that page and try again.
  • I’m not eligible. Is there any way I can still access the challenge?
    There is not. But we regularly have events like this for our Self-Driving Car students, so if you’re excited about self-driving cars you can apply today to enroll in the Udacity Self-Driving Car Program.
  • What will I be challenged to do?
    You'll develop an object detection algorithm similar to the Vehicle Detection project in Term 1 of the Self-Driving Car Nanodegree Program.
  • According to the challenge description, the challenge is similar to the Vehicle Detection project in Term 1, so will my code from that project suffice?
    You can use that same code as a starting point, however, the one critical difference between this challenge and the Vehicle Detection project is that you’ll be required to return pixel-wise labels rather than bounding boxes for detected objects. You’ll also be required to process images at >1 FPS, so you may want to take this opportunity to optimize your code for this challenge.
  • How will participants be evaluated?
    Participants will run their algorithm against an autograder within a classroom workspace and it will be scored for accuracy against a hidden test dataset. Once you’re satisfied with your result you can publish it to the leaderboard for ranking.
  • How does the scoring work?
    Your algorithm will be scored for speed and accuracy. To be eligible for ranking on the leaderboard, your solution must run at greater than 1 frame per second (FPS) in the workspace environment in the classroom.
  • Can I submit more than once?
    You can make as many submissions as you like. However, only your best result will be recorded and ranked on the leaderboard.
  • What are the deadlines for the challenge?
    We will accept submissions until Friday, June 1, 2018 at 5PM PST. Winners will be announced via email the following week.
  • May I enter as a team?
    No, entries are limited to individual submissions. You can feel free to collaborate with others as you like, but only one individual will be associated with each submission, and only that individual will be eligible to earn a job interview.
  • Will my participation be publicly announced?
    No, you will have the opportunity to choose a username for the leaderboard (or be anonymous). The leaderboard will be visible to all students enrolled in or graduated from the Self-Driving Car Program, as well as Udacity staff.
  • Where can I find the official rules for the challenge?
    You can find our complete Official Rules here.
  • Who can I contact with questions or problems related to the Lyft Hiring Challenge?
    Please submit all questions in the Self-Driving Car community's Lyft Challenge Slack channel.