Senior Applied Scientist, Multimodal AI — Seattle, WA
Зарплата не указана
ОткликнутьсяCompany DescriptionOmnidya is a data-focused insurtech company that leverages AI to transform how individuals and businesses experience insurance. The company focuses on using advanced analytics and machine learning to improve decision-making, streamline processes, and deliver more personalized insurance solutions. Team members collaborate across technology, data, and business functions to create products that make insurance more transparent and accessible. Joining 92wBUG2 offers the opportunity to contribute to meaningful innovations in the insurance industry while working with modern AI-driven tools and methods.
Role DescriptionThe Senior Applied Scientist, Multimodal AI is a full-time role based in Seattle, WA with a hybrid work arrangement. In this role, the scientist will build AI that turns driving video, sensor data, images and documents into products customers use, owning models through their full lifecycle: from data and research through training, evaluation and launch, to improving them in production. Key work includes developing a proprietary vision model trained on pUMtWUU'N driving data; building simulation of drivers, vehicles and other road agents to generate scenarios, test models and understand risk; and applying vision-language and language models to explain events and support human decisions. Daily responsibilities include designing datasets and experiments, training and fine-tuning models, building evaluations that show when a model is ready, and deciding with evidence when to build, adapt or buy. The scientist will set technical direction for the AI team as it grows and mentor the engineers and scientists on it. Collaboration with engineering, product and insurance partners, and clear communication of what a model can and cannot do, will be essential to success in this position.
Qualifications• MS or PhD in computer science, machine learning, robotics, computer vision, electrical engineering or a related field, or equivalent industry experience
• 6+ years building and shipping ML models in computer vision, robotics or multimodal AI, including at least one taken from early research to a product in use
• Strong expertise in modern vision models, including training or pretraining on large video or image datasets
• Hands-on experience fine-tuning and evaluating vision-language or language models
• Experience with simulation or behavior modeling, such as agent or driver models, scenario generation, synthetic data or sim-to-real transfer
• Strong dataset skills: designing, curating and labeling data, and knowing where it falls short
• Rigorous evaluation habits: well-built test sets, the right metrics for the cases that matter, and honest reporting of uncertainty
• Fluency in a modern deep learning framework, and code others can build on
• Experience working with engineers to put models into products and improve them from real-world feedback
• A track record of technical leadership and mentoring scientists or engineers
• Clear written and verbal communication with technical and non-technical audiences
• Experience with autonomous driving, robot perception, driver behavior, regulated industries such as insurance, or published research is a plus