LLM Research Scientist (Pre-training & Computer Vision & Adversarial Robustness)

Advance empirical machine learning research across computer vision and language modeling. This role focuses on training, improving, evaluating, and deploying deep learning systems within real data, compute, model-size, and latency constraints.

• Work on well-scoped, open-ended empirical machine learning research problems.
• Train image classifiers and generative image models from scratch, and fine-tune open-weight language models.
• Improve model performance when data, compute, and model-size budgets are limited.
• Strengthen models against adversarial inputs and adversarial conversations.
• Compress models to satisfy strict size and latency requirements while maintaining accuracy.
• Diagnose and resolve model-training issues.

Bring strong expertise in one or more of the following areas:

• Adversarial robustness: adversarial training for image classifiers, including PGD-based training and TRADES; robust-accuracy evaluation under standard threat models such as L∞ attacks and AutoAttack; avoiding gradient-masking pitfalls; and managing robustness-accuracy trade-offs and robust overfitting.
• Efficient computer vision: end-to-end image-classifier training, particularly fine-grained recognition with many visually similar classes and few examples per class; quantization, pruning, and knowledge distillation; and deployment under hard on-device, edge, or embedded size and latency budgets.
• Generative image modeling: training diffusion models, GANs, VAEs, or flow-based models from scratch; iterating against sample-quality measures such as FID; and applying training-efficiency techniques to build strong generators quickly with small parameter counts.
• LLM post-training and behavioral robustness: supervised fine-tuning and preference optimization, including DPO, RLHF, or RLAIF, for open-weight language models; creating datasets through synthetic generation, noisy or weak supervision, and rejection sampling; shaping multi-turn conversational behavior, including resistance to persuasion and sycophancy, calibrated confidence, and appropriate acceptance of corrections; and targeted alignment-style fine-tuning that preserves general capabilities.
• Multilingual pre-training: training multilingual or low-resource-language models from scratch; tokenizer design across scripts and typologically diverse languages; and balancing highly unequal language-data volumes using approaches such as sampling temperatures and cross-lingual transfer in data-constrained settings.
• Additional relevant experience: scaling laws, training-efficiency research, curriculum learning, data ordering, benchmark construction, contamination control, statistically sound model comparisons, uncertainty estimation, model calibration, data augmentation, or synthetic data for robustness.

You should have 3+ years of machine learning research experience, with PhD research counting toward this requirement, and strong experience with PyTorch, JAX, TensorFlow, or a comparable ML framework. A degree from a top-100 university, experience at a FAANG or comparable AI company, or an equivalent research record through publications or impactful open-source contributions is required.

• Remote, hourly engagement.
• Flexible, project-based work.
• Collaborate with leading AI researchers on challenging, high-impact research projects.

$100 to $120 per hour.