Research Engineer, Pretraining, DeepMind

Google is hiring a Research Engineer, Pretraining for DeepMind in Grand Central, Manhattan, US. This role focuses on designing inference‑optimized models and improving efficiency of large transformer systems.

Key Responsibilities

• Design, develop, and deploy inference‑optimized deep learning models for large‑scale transformer architectures.
• Collaborate with product teams to translate research breakthroughs into production‑ready solutions.
• Conduct applied research to enhance model quality, training speed, and serving efficiency.
• Write high‑performance code, perform rigorous experiments, and publish findings internally.
• Mentor junior engineers and contribute to the technical roadmap of DeepMind’s pretraining initiatives.

Required Qualifications

• Bachelor’s degree in Computer Science, Mathematics, Applied Statistics, Machine Learning, or a related field, or equivalent practical experience.
• At least 5 years of hands‑on experience designing inference‑optimized models.
• Proven experience working closely with product teams to deliver AI solutions.

Preferred Qualifications

• Track record of applied research that improves large transformer training or serving efficiency.
• Experience publishing research in top AI conferences or journals.
• Strong programming skills in Python, C++, and TensorFlow/PyTorch.

Compensation

Salary: $209,568.86 per year (fixed).

Frequently Asked Questions

What type of projects will I work on as a Research Engineer at DeepMind?

You will develop and optimize large transformer models, improve training and serving efficiency, and translate research into production AI solutions.

Do I need a PhD for this role?

A bachelor’s degree in a relevant field and 5+ years of experience with inference‑optimized model design are required; a PhD is not mandatory.

Is the salary fixed or does it include bonuses?

The listed salary is $209,568.86 per year; additional compensation details are discussed during the interview process.