Scale is growing rapidly, and joining the Global International Public Sector team is an opportunity to work on one of the most rapidly expanding teams at Scale. This team is responsible for generating, executing, and fostering Scale’s work outside of the United States. There are three core types of work involved:

• Building custom LLMs
• Providing high-quality training data for research institutions building LLMs from scratch
• Partnerships, upskilling, and advisory

As the first MLRE hire on the team, you will partner closely with premier research institutions in the Middle East that are building Arabic-native LLMs to do cutting-edge multilingual research, publish new advancements in ML and multilingual LLMs, and more. Additionally, you will be responsible for scoping and ideating on the optimal ways to approach custom LLM solutions for various use cases. 

You will be involved end-to-end in the research process and potentially be embedded onsite with these institutions’ researchers. If you are excited about shaping the future of the data-centric AI movement, we would love to hear from you!

You will: 

• Publish cutting-edge research in building Arabic / multilingual LLMs
• Ideate on the best and most useful ways to evaluate quality for Arabic LLMs specifically (especially dialectic and safety) 
• Scope custom LLM solutions and approaches to solve real-world problems for our public sector customers
• Quickly prototype and build new deep learning models targeted at frontier capabilities

You have:

• At least 3 to 5 years of model training, deployment and maintenance experience in a production environment
• Strong skills in NLP, LLMs and deep learning 
• Published research in areas of machine learning at major conferences (NeurIPS, ICML, EMNLP, CVPR, etc.) and/or journals
• Native or highly proficient fluency in speaking, writing, and reading Arabic

Nice to haves:

• Experience in dealing with large scale AI problems, ideally in the generative-AI field
• Demonstrated expertise in large vision-language models for diverse real-world applications, e.g. classification, detection, question-answering, etc. 
• Strong high-level programming skills (e.g., Python), frameworks and tools such as DeepSpeed, Pytorch lightning, Kuberflow, TensorFlow, etc. 
• Strong written and verbal communication skills to operate in a cross functional team