Ai Engineer • Mid-Level

Who you are

You have over 3 years of experience in Deep Learning or NLP, and you are skilled in designing and optimizing LLM training pipelines. Your programming expertise in Python and PyTorch allows you to create and evaluate models effectively, ensuring high performance and reliability.

You possess a strong understanding of reinforcement learning techniques, including Direct Alignment Algorithms and methods like PPO, GRPO, and DPO. Your experience with GPU-based training and troubleshooting equips you to handle complex model training scenarios efficiently.

Collaboration is key for you; you thrive in team environments and can communicate effectively with colleagues to achieve common goals. You are also familiar with distributed learning frameworks such as Slurm, DDP, or Horovod, which enhances your ability to work on large-scale projects.

Desirable

You have published papers related to Deep Learning or NLP and may hold a Master's or PhD degree in a relevant field. Experience presenting at major conferences like ACL, EMNLP, or NeurIPS is a plus, showcasing your commitment to advancing the field.

Your familiarity with Docker and Kubernetes allows you to manage containerized applications effectively, and you have experience designing and managing training pipelines using GPU clusters. You understand the nuances of post-training processes and can implement supervised fine-tuning and parameter-efficient fine-tuning techniques.

What you'll do

In this role, you will design and implement LLM training pipelines that are efficient and capable of producing high-quality generative language models for real-world applications. You will continuously explore new methodologies to enhance the quality of the models and ensure they meet user needs.

Your responsibilities will include improving the efficiency of the LLM training process, utilizing reinforcement learning algorithms to optimize model performance. You will work on enhancing the accuracy and stability of generated outputs while preventing reward hacking and designing self-refining learning structures.

You will also develop foundational models that can integrate with external knowledge and APIs, selecting the necessary tools based on the type of directives provided. Your contributions will directly impact the quality of services released to users, making your role crucial in the development of innovative AI solutions.

What we offer

At 42dot, you will be part of a dynamic team that values innovation and collaboration. We provide a supportive environment where you can grow your skills and contribute to cutting-edge projects in AI. You will have opportunities to engage in continuous learning and development, ensuring you stay at the forefront of technology.

We encourage you to apply even if your experience doesn't match every requirement. Join us in shaping the future of AI and making a meaningful impact in the industry.