Research Scientist, Optimization (PhD)
Meta

As a research scientist in this role, your primary objective will be to develop creative solutions that optimize computational efficiency for training and on-device inference of foundational vision and language models for use cases in VR, AR and mobile devices. Models of interest range from foundational models like Segment Anything, Anything 3D, to multi-modal LLMs and diffusion models. Optimizations can range all the way from kernel level optimizations, efficient building blocks for models, improved model architectures and pre-training techniques. Strong expertise in building efficient models is preferred for this position.

Research Scientist, Optimization (PhD) Responsibilities:

• Develop model optimization algorithms and infrastructure to enable deployment of models in resource-constrained settings, such as mobile phones and AR/VR and MR devices.
• Optimize models on hardware to achieve the best performance given various real time latency and power constraints.
• Develop novel techniques to improve training efficiency and data efficiency.
• Lead and contribute to cutting-edge research that results in industry-leading tech demos and/or publications.
• Collaborate with cross-disciplinary research and engineering teams.

Minimum Qualifications:

• Currently has, or is in the process of obtaining, a PhD degree in Machine Learning (ML) Systems, Natural Language Processing (NLP), Machine Learning, Artificial Intelligence, relevant technical field, or equivalent practical experience. Degree must be completed prior to joining Meta.
• Currently has, or is in the process of obtaining a Bachelor's degree in Computer Science, Computer Engineering, relevant technical field, or equivalent practical experience. Degree must be completed prior to joining Meta.
• Hands-on experience in deep learning algorithms and techniques, e.g., transformers, convolutional neural networks, quantization and/or related areas.
• 3+ years experience with deep learning software libraries such as PyTorch or TensorFlow.
• Demonstrated software engineering experience within a complex codebase via an internship, work experience, coding competitions, or widely used contributions to open source repositories such as GitHub.
• Must obtain work authorization in country of employment at the time of hire, and maintain ongoing work authorization during employment.

Preferred Qualifications:

• Contribution (as the first author) to publications on machine learning at leading workshops or conferences such as ICLR, NeurIPS, ACL, EMNLP, ICML, MLSys, KDD, OSDI etc.
• Experience, including PhD research, in related fields of improving the training and inference efficiency, data efficient learning, domain adaptation and semi-supervised learning.
• Experience, including PhD research, in kernel development for both GPU and mobile devices.
• Experience, including PhD research, in in related fields of efficient deep learning, training acceleration, quantization, model/accelerator co-design, and co-optimization.
• Interpersonal experience: cross-group and cross-culture collaboration

About Meta: