Adversarial ML Scientist: Robustness & Efficient Models

Mercor; San Diego (CA); On-site; USD 150,000 - 210,000; Full time

Job summary

Obsidian is seeking experienced machine learning researchers to advance end-to-end deep learning across vision and language. You will tackle well-scoped empirical problems and push state-of-the-art models.

You will train image classifiers and generative models, optimize data and compute budgets, and focus on robustness and efficiency. The role offers flexible project-based work with high-impact opportunities.

Qualifications

• 3+ years of machine learning research experience (PhD counts toward this requirement).
• Strong experience with PyTorch, JAX, TensorFlow, or similar ML frameworks.
• Degree from a top-100 university, experience at a FAANG or comparable AI company, or an equivalent research track record through publications or impactful open-source contributions.

Responsibilities

• Train image classifiers and generative image models end-to-end, and fine-tune open-weight language models.
• Get the most out of limited data, compute, and model-size budgets.
• Make models robust — to adversarial inputs and to adversarial conversations.
• Compress models to meet hard size and latency constraints without sacrificing accuracy.
• Diagnose and resolve training issues.

Skills

PyTorch/JAX/TensorFlow

Adversarial robustness

Model compression

Generative modeling

Multilingual training

Education

PhD or equivalent research track record

Tools

PyTorch

JAX

TensorFlow

Job description

Obsidian is seeking experienced machine learning researchers to advance end-to-end deep learning across vision and language. You will tackle well-scoped empirical problems and push state-of-the-art models.

You will train image classifiers and generative models, optimize data and compute budgets, and focus on robustness and efficiency. The role offers flexible project-based work with high-impact opportunities.