Company Description:

MosaicML is a deep learning startup with a mission to make ML training more efficient for everyone through fundamental innovations in algorithms, systems, and platforms. We believe that large scale training should be available beyond the well-resourced companies, and bridging the gap between research and industry is core to our success. Our products will enable our customers to train the best neural networks efficiently as possible within given time, cost, or other resource constraints – and to do so with a great user experience.

Job description:

Develop methods for efficient neural network training. You will survey ideas in the literature and develop ideas of your own to change the way neural networks are trained to improve efficiency. This involves exacting scientific inquiry, rigorous empirical analysis of large-scale experiments, and building high-quality research artifacts.

Systematize knowledge and adjudicate scientific truth. At MosaicML, we are focused on transforming scientific knowledge into practical efficiency improvements. To do so, we navigate the messy machine learning literature, determine what holds up under scrutiny, and use that knowledge to build better models.

Advance the frontier of deep learning. You will drive ambitious research projects that push the limits of existing technology and explore new approaches that go beyond the state of the art.

Love our customers. Our goal is to make our customers successful when they train large deep learning models. We seek to encode our scientific expertise in our tools for the benefit of our customers. We love our customers, and we expect you to love them too!

What we’re looking for:

The strongest candidates will have experience with at least one of the following:

Experience training large models. We’re looking to hire researcher scientists and research engineers who have experience training modern neural networks for computer vision, natural language processing, and multimodal settings. Ideally, you will have experience training at large scales (100M+ parameters, and ideally 1B+ parameters) and conducting multi-node training. 

Extensive experience with NLP for deep learning. We are looking to hire research scientists who specialize in natural language processing.

Experience with data preparation and data quality for deep learning. We are looking to hire research scientists who have experience cleaning and curating large-scale data corpora (1B-100B examples) and using those corpora for deep learning.

Experience with deep generative model evaluation and improvement. We are looking to hire research scientists with experience evaluating generative models and using those insights to improve the models. 

A PhD is NOT required for this role. We are open to hiring candidates with bachelor’s and master’s degrees and to new graduates. We are open to hiring candidates who are currently in "research engineer" roles at other companies.

Your responsibilities:

• Keeping up to date with the research literature and thinking beyond the current state of the art.
• Developing and implementing methods that improve the efficiency and efficacy of deep learning.
• Rigorously evaluating these methods and communicating the results of your findings.

Your qualifications and qualities:

• Proficiency with the fundamentals of deep learning.
• Proficient software engineering skills and proficiency with PyTorch.
• Knowledge of the systems aspects of how neural networks train and the resources used in the process of doing so.
• Research experience in deep learning.

Additional Notes:

• US work authorization required
• Salary Range: $145K - $295K
• Also includes equity (stock options) and benefits