We’re announcing Vision: our enhanced, production-ready urban perception system for Level 3 autonomous driving. Our updated vision-first system now includes BEV fusion functionality, ISO 26262 ASIL-B(D) certification, ASPICE Level 2 assessment, and

(4/4) Validated for mass production and fully compatible with the end-to-end Driver path planning stack, Vision enables reduced validation effort and increased interpretability, streamlining deployment of full-stack AI software.

(3/4) Modular by design, Vision is optimized for deployment on leading hardware platforms including Nvidia, Qualcomm, TI, and Ambarella.

(2/4) New features include bird’s-eye view (BEV) generation from multi-camera input, ISO 26262 ASIL-B(D) certified components, and an ASPICE Level 2 assessment. These enhancements support more accurate downstream planning and confirm the system’s readiness for integration into.

(1/4) Vision builds on our surround-view architecture and is trained on real-world data using our proprietary Deep Teaching™ methodology. It uses camera input and scales across diverse geographies, road geometries, and traffic behaviors.

3/.Driver is fully compatible with our production-grade vision perception stack, enabling a modular, interpretable, and scalable self-driving system across geographies, vehicle platforms, and driving conditions. Learn more:

2/.To showcase the model in a realistic, dynamic environment, we ran Driver in closed-loop simulation using the open-source @carlasimulator. GenSim-2, our generative foundation model, re-rendered the simulated scenes to produce realistic camera.

1/.Designed for Level 2–4 autonomy, Driver handles complex driving scenarios like intersections, turns, obstacle avoidance, and vehicle cut-ins. These human-like behaviors emerge from end-to-end learning, powered by Deep Teaching™—not from hand.

Introducing Driver — a transformer-based, real-time, vision-only path prediction model for highway and urban driving. It predicts the vehicle’s future path using only camera-based perception—no Lidar, HD maps, or additional sensors required. We deployed

Watch to learn the journey of building from a decade in mathematics and two years of pure R&D, to securing over $100M in funding and developing first-of-its-kind AI models for autonomous driving in partnership with global automakers. Watch on YouTube:

Our CEO & Founder @VVoroninski had a great conversation with @DanielDarling on his podcast 5 Year Frontier about the future of autonomous driving and robotics!. They discussed simulating physical environments, helping car makers leapfrog Tesla, real world AI models, the

This week, our CEO & Founder @VVoroninski discusses with @IEEESpectrum's Eliza Strickland one of autonomous driving's biggest challenges: preparing for rare, unexpected scenarios. In the interview, he shares how s foundation models like WorldGen and

#artificialntelligence #AI #Mining.

In the recent interview with @SAEIntl, s CEO Vladislav Voroninski discusses how our AI technology—originally developed for autonomous driving—is demonstrating its applicability in mining. The same Deep Teaching™-powered perception systems that enable

We’re excited to share that has achieved Automotive SPICE (ASPICE) Capability Level 2!. This achievement underscores our commitment to meeting the rigorous standards required for the volume production of safety-critical software in mass-market road

GenSim-2, s generative AI model with video editing capabilities for autonomous driving, scalably and cost-effectively creates and edits video data by modifying weather and lighting conditions—such as rain, snow, fog, and day-to-night transitions. Powered

Last week at hashtag#CES2025, showcased how we’re redefining autonomous driving with our AI-first solutions. Below are highlights:. 1️⃣ Honda revealed the prototypes of its mass-production vehicles, Honda Zero Saloon and Honda Zero SUV. The Zero Series will

4/4:.GenSim-2 bridges the gap between simulation and real-world testing, accelerating development timelines while reducing costs. Check out the full press release (. Stay tuned for more updates!.

3/4:.Multi-camera support ensures consistency across all perspectives and high realism for surround-view camera stacks.

2/4:.Fully customize object appearances:.- Vehicles: type, color, and condition.- Roads: paved, cracked, wet etc.- Rare objects like guardrails.- Vegetation, buildings, pedestrians, and more.