S
Lead Research Scientist
В офисеПолная занятостьAI & Machine LearningMobile GamingAdTech
Lead an open-ended deep learning research program for an AI platform serving mobile gaming. Define research direction, build models, establish rigorous experiments, and mentor researchers while helping translate validated work into product capabilities.
Обязанности
- Define and evolve the research program, including research questions, scope, priorities, and objectives.
- Design, implement, and train deep learning models, architectures, representations, and learning objectives.
- Establish baselines, guide data collection and dataset design, and develop evaluation methods for capability and generalization assessment.
- Own project progress by identifying obstacles, coordinating collaborators, and deciding when to deepen, change, or stop research directions.
- Provide technical guidance and mentorship to researchers and strengthen their scientific approaches.
- Collaborate with engineering and product teams to turn validated research into usable capabilities.
Требования
- Ph.D. focused on deep learning or equivalent research experience with comparable scientific and technical depth.
- Track record leading substantial deep learning research, including problem formulation, methodology, experimental direction, and outcomes.
- Hands-on experience designing or adapting model architectures and training methods.
- Experience implementing experiments with frameworks such as PyTorch or JAX and diagnosing model behavior.
- Strong research judgment in selecting questions, approaches, experiments, and interpreting evidence.
- Ability to independently drive complex, long-horizon research projects through uncertainty and technical obstacles.
- Technical depth, adaptability, and intellectual independence across challenging research problems.
- Experience providing technical leadership and mentoring other researchers.
Будет плюсом
- Experience with world modeling, video or multimodal representation learning, sequential modeling, imitation learning, reinforcement learning, or human-behavior modeling.