Applied Scientist

What you'll do

• Identify research advances with potential for differentiated product capabilities.
• Manage applied research from problem formulation through prototype implementation and validation.
• Curate and analyze multimodal training data including sampling, filtering, and synthetic generation.
• Design and optimize multimodal architectures like vision encoders and cross-modal attention.
• Build rigorous evaluation pipelines for model quality, robustness, safety, and latency.
• Diagnose model failures through ablation studies and architecture-level experimentation.
• Implement post-training techniques such as supervised fine-tuning and reinforcement learning.
• Develop, debug, and improve research-quality code using Python and PyTorch.

What we're looking for

• Master’s or Ph.D. in computer science, machine learning, artificial intelligence, or a related field, or equivalent practical experience.
• Hands-on experience developing transformer, diffusion, or multimodal architectures through direct training or adaptation of model parameters.
• Strong Python and PyTorch skills for developing, debugging, and improving research-quality model code.
• Experience with end-to-end model development including data preparation, training, evaluation, and failure analysis.
• Experience with post-training techniques such as supervised fine-tuning, preference optimization, reinforcement learning, and alignment.
• Ability to communicate technical decisions, experimental findings, and measurable impact effectively.
• Preferred experience with large-scale or distributed training and compute- or memory-efficient training techniques.
• Preferred experience with inference optimization, synthetic-data generation, and quality control.