Research Engineer, Pre-training

Anthropic is seeking a Research Engineer to join their Pre-training team, focusing on developing the next generation of large language models. This role represents a unique opportunity to work at the intersection of cutting-edge AI research and practical engineering implementation.

The position involves working on critical aspects of AI development, including model architecture, algorithms, data processing, and optimizer development. As a Research Engineer, you'll be responsible for conducting research, implementing solutions, and leading small research projects while collaborating with team members on larger initiatives. The role requires expertise in Python and deep learning frameworks, with a preference for PyTorch experience.

The company offers a competitive salary range of $315,000 - $340,000 USD and provides comprehensive benefits including visa sponsorship, flexible working hours, and generous vacation time. The work environment is hybrid, requiring at least 25% in-office presence at one of their locations in San Francisco, Seattle, or New York City.

Anthropic stands out for its approach to AI research, viewing it as an empirical science similar to physics and biology. The company operates as a single cohesive team focused on large-scale research efforts, prioritizing impact and the development of steerable, trustworthy AI systems. Their research continues important work in areas such as GPT-3, Circuit-Based Interpretability, Multimodal Neurons, and AI Safety.

The ideal candidate will have an advanced degree in Computer Science, Machine Learning, or a related field, strong software engineering skills, and experience with large-scale machine learning systems. They should be comfortable balancing research goals with practical engineering constraints and have excellent communication skills for collaborative work.

Key projects include optimizing novel attention mechanisms, comparing compute efficiency of different Transformer variants, preparing large-scale datasets, scaling distributed training jobs, and creating interactive visualizations of model internals. The role offers an opportunity to contribute to the development of safe, ethical, and powerful artificial intelligence systems while working with a team committed to ensuring AI systems are aligned with human interests.

Anthropic values diversity and strongly encourages applications from candidates of all backgrounds, including those from underrepresented groups in tech. The company operates as a public benefit corporation and offers additional benefits such as equity donation matching and a collaborative office space for team interaction.