Mobileye

My group in Mobileye is looking for experienced algorithms and ML engineers and researchers to work on the most challenging problems of autonomous driving. Our group works closely with Mobileye's CTO, Prof. Shai shalev-shwartz, on the cutting edge of autonomous driving technology. We develop production grade software for our current self driving products and research and implement next generation technologies. Bonus: we get to ride in our autonomous vehicles quite a lot and see our work in action. Positions are available both in our Jerusalem and Ramat Gan offices. Feel free to reach out with questions / discussions and to apply. Also feel free to share with people that might be interested. Algorithms: https://lnkd.in/dcfkpAHg https://lnkd.in/e5yrCzcf ML: https://lnkd.in/eWajqc6k https://lnkd.in/eR_hMecq RL: https://lnkd.in/eg5rDdm7

Three people died trusting a hands-free system. The National Transportation Safety Board (NTSB) just told the whole industry why. The findings don't point to a single broken product. They point to a broken design philosophy, and it's everywhere. Three specific gaps were cited: 1️⃣- No requirements to record crash data during ADAS incidents. 2️⃣- Safety-critical features drivers could disable while hands-free mode was active. 3️⃣- Driver monitoring systems that couldn't detect real distraction. None of these are exotic failures. They're the predictable result of deploying partial automation without formally defining what the system is actually responsible for, and what it isn't. That's the core problem. The industry has been building hands-free systems around a question it hasn't fully answered: what exactly is a dangerous situation, and what is the correct response to it? When those definitions are vague, system design becomes discretionary. Features get tuned for user experience rather than safety boundaries. Monitoring gets optimized for low false-positive rates rather than genuine attention detection. And the human in the loop gets conditioned to trust a system that was never built to deserve that trust unconditionally. This is where the right approach starts upstream. Before a line of code is written, you need formal mathematical definitions of dangerous situations, safe distances, and proper responses - not design guidelines, not best practices, not "no defects found." Formal definitions that can be verified, tested against, and held to. That's the foundation Mobileye 's safety architecture is built on, and it's the reason the NTSB findings read, to me, as a validation of that approach rather than a surprise. The regulatory pressure coming out of this standardized performance requirements, mandatory data recording, driver monitoring standards will eventually land on everyone. The question is whether the industry waits for the mandate or gets ahead of it. What's your read: does this accelerate federal ADAS standards, or does it stall in the same regulatory fog that's kept L2 ungoverned for years? #AutonomousDriving #ADAS #FunctionalSafety #AVPolicy #AunotmousVehicle Opinions are my own.

Hi everyone, My team at Mobileye is currently hiring for a Deep Learning Algorithm Developer role in NSS. There is a preference for candidates with some relevant experience in the field, or a relevant MSc degree. If this sounds interesting to you, or to someone you know, you can apply through my referral link here: https://lnkd.in/eEu6h7R6 You can also check out and apply to other open Mobileye positions through this live referral link: https://lnkd.in/eJ9H-qHW Also, in completely unofficial news, there is an imaginary bonus for anyone who adopts a dog through us, or at least follows the rescue organization I volunteer with on Instagram (the bonus is fake, but please do follow for real 🤭😛): https://lnkd.in/epESGkXb If you know someone who might be a fit, I’d really appreciate a share.