Mobileye izstrādā progresīvas autovadītāja palīdzības sistēmas un autonomās braukšanas tehnoloģijas automobiļu nozarei. Uzņēmums apvieno datorredzi, mākslīgo intelektu, paša izstrādātu aparatūru un programmatūru, lai atbalstītu transportlīdzekļu drošību risinājumos, kas aptver gan autovadītāja palīdzību, gan autonomu braukšanu. Uzņēmuma sistēmas ir ieviestas vairāk nekā 190 miljonos transportlīdzekļu, apliecinot Mobileye nozīmi uz redzi balstītu automobiļu tehnoloģiju ieviešanā plaša mēroga ražošanā.
Applied AI Researcher, Visual GenAI at Mobileye (Hybrid, Israel)
Advance visual GenAI, diffusion models, and computer vision for Mobileye’s autonomous driving technology. Build synthetic-data solutions using proprietary datasets and large-scale computing infrastructure.
Apraksts
- Lead the full development lifecycle, from training generative models to adapting state-of-the-art architectures for real-world autonomous driving applications.
- Research, assess, and enhance GenAI and computer vision models through scalable, clean Python implementations.
- Partner with AI researchers and perception teams to apply synthetic data in real-world AI systems.
- Track emerging developments in artificial intelligence, generative modeling, and computer vision.
- Create GenAI solutions with large proprietary datasets and high-scale computing infrastructure for autonomous vehicle applications.
Prasības
- Master’s or doctoral degree in Computer Science, Electrical Engineering, or a related discipline.
- At least three years of industry experience in deep learning and computer vision.
- Advanced proficiency in Python and PyTorch.
- Strong capabilities in problem solving, research, and analytical reasoning.
- Practical experience with generative AI techniques such as diffusion models and GANs.
- Experience with vision-language models and multimodal learning.
- Publications at leading conferences such as CVPR, NeurIPS, or ECCV.
- Hands-on experience with Linux, AWS, Docker, and virtual environments.
- Knowledge of traditional computer vision methods.