LLM Research Scientist – Pre-training, Computer Vision, Adversarial Robustness

This is a fully remote position, open to applicants in United States.

📋 Description

• Develop image classifiers and generative image models from the ground up.

• Refine open-weight language models through fine-tuning.

• Enhance model performance within constraints of data, computational resources, and model size.

• Strengthen model resilience against adversarial inputs and conversations.

• Reduce model size to comply with latency requirements while maintaining accuracy.

• Identify and address issues during the training process.

• Engage in empirical, exploratory machine learning research.

• Partner with premier AI researchers on influential projects.

• Participate in initiatives focused on training and enhancing AI systems for top AI laboratories and enterprises.

⛳️ Requirements

• Minimum of 3 years of experience in machine learning research; PhD research is applicable towards this criterion.

• Proficient in PyTorch, JAX, TensorFlow, or similar machine learning frameworks.

• Degree from a top-100 university, experience at a FAANG or equivalent AI organization, or a comparable research portfolio demonstrated through publications or notable open-source contributions.

• Specialized knowledge in one or more areas such as adversarial robustness, efficient computer vision, generative image modeling, post-training for LLMs, behavioral robustness, or multilingual pre-training.

• Background in adversarial training of image classifiers, robust accuracy assessment, and navigating robustness–accuracy trade-offs.

• Experience in end-to-end training of image classifiers, model compression, and deployment under strict size or latency constraints.

• Familiarity with training diffusion models, GANs, VAEs, or flow-based generative models from inception.

• Experience with supervised fine-tuning and preference optimization of open-weight language models.

• Knowledge in shaping conversational behavior and alignment-style fine-tuning.

• Experience in training multilingual or low-resource language models from scratch and developing tokenizers across various scripts.

• Additional expertise in scaling laws, curriculum learning, model evaluation, uncertainty estimation, calibration, or synthetic data is advantageous.

• Must be an independent contractor.

• Mercor is currently unable to sponsor H1-B or STEM OPT candidates.

🏝️ Benefits

• Flexible, project-based work opportunities.

• Competitive pay structure.

• Fully remote work environment.

• Freedom to establish your own work schedule.

• Weekly payments processed via Stripe or Wise.

• Project durations may be extended, shortened, or concluded early based on needs and performance.

• Reasonable accommodations available upon request.

• Referral compensation of up to $480 per successful referral.