AI Foundation Researcher: Large Language Models & VLMs

Jobtailor; California (MO); On-site; USD 180,000 - 240,000; Full time

Job summary

Jobtailor is seeking a research scientist to advance AI foundation models, leading end-to-end development from design to deployment. You will collaborate with data scientists, software engineers, and product managers to translate cutting-edge AI research into customer-focused solutions.

Ideal candidates hold a PhD or MS with strong industrial NLP or ML research experience, with a track record of publications or large-scale model work and comfort with cloud computing platforms.

Qualifications

• PhD in Electrical Engineering, Computer Engineering, Computer Science, AI, Mathematics, or related fields; degree may be obtained on or before the scheduled start date.
• MS in Electrical Engineering, Computer Engineering, Computer Science, AI, Mathematics, or related fields plus 2 years of experience in Applied Research.
• Deep understanding of AI methodology foundations.
• Experience building large deep learning models for language, images, events, or graphs.
• Expertise in one or more of: training optimization, self-supervised learning, robustness, explainability, or RLHF.
• Track record delivering models at scale in training data and inference volumes.
• Experience delivering libraries, platform-level code, or solution-level code to existing products.
• Track record of high-quality machine-learning ideas, such as first-author publications or projects.
• Ability to own and pursue a research agenda and autonomously carry out long-running projects.
• Hands-on experience developing AI foundation models and solutions using open-source tools and cloud computing platforms.
• Preferred: PhD in a related technical field.
• Preferred: NLP focus or Master's with 5 years of industrial NLP research experience.
• Preferred: publications related to pre-training large language models, deep learning theory, or major ML conferences.
• Preferred: experience training a large language model from scratch with 10B+ parameters and 500B+ tokens.
• Preferred: experience optimizing training for a 10B+ model.
• Preferred: deep-learning algorithmic or optimizer design knowledge.
• Preferred: compiler design experience.
• Preferred: knowledge of transfer learning, model adaptation, and model guidance.
• Preferred: experience deploying a fine-tuned large language model.
• Capital One will consider sponsoring a new qualified applicant for employment authorization.

Responsibilities

• Partner with data scientists, software engineers, machine learning engineers, and product managers to deliver AI-powered products.
• Use PyTorch, AWS Ultraclusters, Hugging Face, Lightning, VectorDBs, and other technologies to uncover insights from large numeric and textual datasets.
• Build AI foundation models through design, training, evaluation, validation, and implementation.
• Conduct applied research and advance emerging AI developments into customer experiences.
• Translate complex technical work into tangible business goals.
• Research and evaluate emerging technologies and state-of-the-art AI methods.
• Define and solve ambiguous problems and improve the status quo with stakeholders.
• Contribute to talent development.

Skills

Deep Learning

NLP

AI Methodology Foundations

Training Optimization

Self-Supervised Learning

Robustness

Explainability

RLHF

Large Language Model Training

Education

PhD in Electrical Engineering/Computer Engineering/CS/AI/Mathematics

MS in EE/CE/CS/AI/Math + 2 years Applied Research

Tools

PyTorch

AWS Ultraclusters

Hugging Face

Lightning

VectorDBs

Open-Source Tools

Cloud Computing Platforms

Compiler Design

Job description

Jobtailor is seeking a research scientist to advance AI foundation models, leading end-to-end development from design to deployment. You will collaborate with data scientists, software engineers, and product managers to translate cutting-edge AI research into customer-focused solutions.

Ideal candidates hold a PhD or MS with strong industrial NLP or ML research experience, with a track record of publications or large-scale model work and comfort with cloud computing platforms.