Mobileye

The threat model at Mobileye starts with physics: a vehicle is a two-ton machine moving at highway speed, and a compromised sensor stack or a poisoned training pipeline isn't a data breach - it's a potential fatality. Founded in 1999 and headquartered in Jerusalem, the company builds the compute substrate (EyeQ chips) and full perception-to-decision stack that underpins advanced driver-assistance systems and fully autonomous platforms. More than 230 million EyeQ chips have shipped globally as of 2025, embedded in vehicles across multiple continents. That's a massive, safety-critical attack surface baked into production fleets worldwide.

Engineering here spans purpose-built silicon, computer vision algorithms, and machine learning models that must be provably robust against adversarial inputs - weather, sensor degradation, and deliberate manipulation. The Mobileye Drive platform extends that stack to full autonomy, which means the security team isn't just defending IT assets; they're hardening systems where failure modes are measured in seconds and meters, not downtime and logs. The company's culture positions engineering safety as a foundational property, not a compliance checkbox, and empowers individual contributors to drive outsized impact.

Mobileye also operates at the intersection of autonomous mobility and physical AI, with work extending into humanoid robotics. The technical domains at play - AI, computer vision, hardware and software development - create a security challenge that's cross-layer by nature: silicon-level trust anchors, firmware integrity, ML model provenance, over-the-air update pipelines, and vehicle-to-cloud communication all need to hold up at scale. If you're looking to do security work where the consequence model is tangible and the systems are genuinely complex, this is one of the few places where that's not an abstraction.