Hands-on ML research on well-scoped, open-ended problems across vision and language. You train and improve deep learning models end to end, under tight data, compute and size budgets, in flexible project-based work.
What you will do
- Train image classifiers and generative image models from scratch, and fine-tune open-weight language models.
- Get the best results from limited data, compute and model size.
- Make models robust to adversarial inputs and adversarial conversations.
- Compress models to meet hard size and latency limits without losing accuracy.
- Find and fix training problems.
What you need
- 3+ years of ML research experience (PhD research counts).
- Strong experience with PyTorch, JAX, TensorFlow or a similar framework.
- A degree from a top-100 university, experience at a FAANG or similar AI company, or an equal research record through publications or open-source work.
- Strong expertise in at least one area: adversarial robustness, efficient computer vision and model compression, generative image models (diffusion, GANs, VAEs, flows), LLM post-training and behavioral robustness (SFT, DPO, RLHF, RLAIF), or multilingual pre-training and tokenizer design.
A plus: scaling laws, curriculum learning, benchmark construction and contamination control, uncertainty estimation and calibration, or data augmentation and synthetic data.
How pay works
The listing gives an hourly rate of $100-120/hr for flexible, project-based work.
Apply (opens in a new tab)