🕒 17 days ago
Research and develop inference algorithms and model architectures for Cerebras hardware, advancing speculative decoding, model compression, sparse attention, and other techniques to improve inference quality, latency, and throughput.
Member of Technical Staff - World Models
📅 Sep 7
Machine Learning Foundation ModelsPytorchDistributed TrainingGenerative Modeling
📜 Description
- Design, train, and scale generative and predictive models across visual and spatial modalities.
- Develop multimodal learning methods using diverse data sources and conditioning signals.
- Improve consistency, robustness, and generalization in models of complex temporal and spatial structures.
- Explore post-training and optimization techniques to improve model quality, controllability, and computational efficiency.
- Develop rigorous evaluations and run controlled experiments to guide modeling and data decisions.
- Collaborate with data, systems, and robotics engineers to turn research ideas into reliable capabilities.
🛠️ Requirements
- PhD degree in Computer Science, Engineering, or a related technical field, or equivalent experience.
- Experience training machine learning foundation models end to end, with strong PyTorch skills and hands-on experience in distributed training.
- Ability to independently design, execute, and analyze machine learning experiments, from hypothesis through ablation to conclusion.
- Experience developing evaluations that reveal model limitations and inform research decisions.
- Strong publication record or substantial contributions to research on image, video, 3D, or multimodal models.
- Experience with learned representations, compression, or tokenization for high-dimensional data.
- Experience with conditional generation, temporal consistency, or long-context modeling.
- Experience with reinforcement learning, model-based learning, or embodied AI.
- Experience with post-training, distillation, or efficient inference for generative models.
- Experience working with large-scale, heterogeneous visual or sensor datasets.
Full job description
Member of Technical Staff – World Models
About Us
Veeda AI is building the next generation of multimodal foundation world models for Physical AI. We're a small, fast-moving team of engineers and researchers from leading AI labs, tackling some of the most challenging problems at the intersection of AI, robotics, and embodied intelligence. If you're excited about pushing the boundaries of what's possible with Physical AI, you'll have the opportunity to make an outsized impact from day one.
Responsibilities
Model Research & Development: Design, train, and scale generative and predictive models across visual and spatial modalities.
Multimodal Learning & Control: Develop methods for learning from diverse data sources and incorporating conditioning signals into model behavior.
Sequence Modeling: Improve consistency, robustness, and generalization in models of complex temporal and spatial structure.
Post-Training & Efficiency: Explore post-training and model optimization techniques to improve quality, controllability, and computational efficiency.
Evaluation & Experimentation: Develop rigorous evaluations, conduct controlled experiments, and use findings to guide modeling and data decisions.
Research Collaboration: Work closely with data, systems, and robotics engineers to translate research ideas into reliable capabilities.
Requirements
PhD degree in Computer Science, Engineering, or a related technical field, or equivalent experience.
Experience training machine learning foundation models end to end, with strong PyTorch skills and hands-on experience in distributed training.
Ability to independently design, execute, and analyze machine learning experiments, from hypothesis through ablation to conclusion.
Experience developing evaluations that reveal model limitations and inform research decisions.
Nice to Have
Strong publication record or substantial contributions to research on image, video, 3D, or multimodal models.
Experience with learned representations, compression, or tokenization for high-dimensional data.
Experience with conditional generation, temporal consistency, or long-context modeling.
Experience with reinforcement learning, model-based learning, or embodied AI.
Experience with post-training, distillation, or efficient inference for generative models.
Experience working with large-scale, heterogeneous visual or sensor datasets.
Contributions to open-source machine learning projects or research infrastructure.
Similar jobs
Search more Research Scientist jobs
📅 Sep 9
Own the end-to-end mid-frequency crypto research agenda, developing and deploying trading models across derivatives and spot markets while shaping research direction, monitoring live performance, and mentoring colleagues.
🕒 yesterday
Research and improve voice AI models by turning customer problems into measurable goals, designing evaluations, curating data, fine-tuning models, and shipping production-ready solutions.
Forward Deployed Applied Scientist II (Agentic AI)
🕒 4 days ago
DigitalOcean👥 1001 - 5000 employees🏢 Computer Software
🕒 4 days ago
Design and deploy production-ready agentic AI systems for customer challenges, combining applied research, multi-agent workflows, reliability evaluation, and direct collaboration with customers and product teams.
🕒 5 days ago
Lead research into identity verification and fraud prevention, developing and evaluating methods to detect manipulated media and attacks, assess model performance, and turn findings into production defenses.