Scot Lewis Associates Ltd
16/09/2024
Full time
Senior LLM Engineer I'm currently working with an exciting, forward-thinking company that's looking to bring on a Senior Large Language Model (LLM) Engineer to join their fully remote international team. They're focused on cutting-edge AI technologies, especially in deploying highly realistic AI characters and models at scale. If you're passionate about training and fine-tuning open-source LLMs and working with the latest AI models, this could be a perfect opportunity for you. The Role: You'll be responsible for training open-source LLMs (like Llama and Mistral) and fine-tuning them for immersive, high-quality chat experiences. Build, curate, and manage datasets, including leveraging models like GPT-4 to improve smaller LLMs. Help deploy these models at scale using tools like Hugging Face's Text Generation Inference (TGI). Collaborate with a remote team using asynchronous communication (Slack, video meetings). Tech Stack You'll Be Using: LLMs & Frameworks: Llama 3, Mistral, Axolotl, HF Transformers, PyTorch, NumPy Infrastructure: Linux, Docker, AWS, Kubernetes, Cloud GPU Providers Inference: Hugging Face TGI What You'll Need: At least 1 year of experience training open-source LLMs. Strong understanding of fine-tuning techniques (SFT, LoRA/qLoRA, RLHF). Python 3.x skills with experience in typing and data validation (Pydantic). Familiarity with HF Transformers, PyTorch, and building/manage datasets. Bonus if you've worked on conversational AI models, NSFW content, or have experience with Axolotl. What's in It for You: A fully remote, flexible working environment. The chance to work with some of the most innovative AI minds globally. Competitive compensation package. This is an excellent opportunity if you're looking for a challenge in the AI space and want to make a real impact with your skills.