Research Scientist – Multimodal Foundation Models & Visual Reasoning

Elorian • Palo Alto, CA

2d

About The Position

The Research Scientist will lead ambitious research across next-generation model architectures, distributed training methodologies, multimodal visual reasoning, reinforcement learning, and data-centric learning. With dedicated large-scale compute clusters and engineering support, the role involves developing first-principles hypotheses and running large-scale empirical experiments. The scientist will work closely with research and engineering colleagues to advance visual intelligence and translate research breakthroughs into state-of-the-art foundation models and real-world capabilities.

The position focuses on frontier research in foundation models, multimodal understanding, complex visual reasoning, and spatial intelligence. It also includes conceiving and building novel architectures, pre-training methods, and scalable post-training techniques, as well as designing and analyzing distributed training experiments. The role will pioneer multimodal data curation and synthetic data approaches, evaluate models rigorously, and collaborate across research and systems engineering to deploy robust capabilities at scale.

Requirements

• Ph.D. or equivalent industry research track record in Machine Learning, Computer Vision, Natural Language Processing, or a related quantitative discipline.
• Rigorous foundational understanding of deep learning principles, transformer and multimodal architectures, and modern large-scale training dynamics.
• Deep proficiency in Python and modern deep learning frameworks such as PyTorch or JAX.
• Demonstrated ability to independently define research agendas, design rigorous empirical experiments, and solve open-ended, frontier-level problems.

Nice To Haves

• Deep theoretical and practical expertise in at least one of foundation models, multimodal reasoning, visual or spatial intelligence, reinforcement learning, or post-training.
• Strong research record, demonstrated through first-author publications at top-tier venues such as NeurIPS, ICML, ICLR, CVPR, ICCV, or ECCV, or key technical contributions to high-profile foundation model releases.
• Hands-on experience training, profiling, and scaling models across large distributed GPU or TPU clusters, using tools such as FSDP, Megatron-LM, DeepSpeed, or Ray.
• Experience developing large-scale multimodal data pipelines, generating synthetic reasoning traces, or designing rigorous evaluation suites for frontier models.

Responsibilities

• Drive research at the intersection of foundation models, multimodal understanding, complex visual reasoning, and spatial intelligence.
• Conceive and build novel model architectures, pre-training methodologies, and scalable post-training techniques, including preference tuning and reasoning-focused reinforcement learning.
• Design, execute, and analyze large-scale distributed training experiments on modern accelerator clusters.
• Develop approaches to multimodal data curation, synthetic data generation, automated filtering, and benchmark evaluation.
• Collaborate with research and systems engineering teams to turn algorithmic breakthroughs into robust, scalable foundation model capabilities.

What This Job Offers

• Job Type: Full-time
• Career Level: Senior
• Education Level: Ph.D. or professional degree