remote
AI/ML LLM Engineer - https //ladybirdltd.com
LLM Engineer
Develop and fine‑tune large language models for an AI‑powered NHS receptionist, integrating OpenAI GPT for early testing and building a custom LLaMA 3 model with synthetic data, LoRA/SFT, RLHF/DPO, and SageMaker deployment.
About the role
Key Responsibilities
- Integrate OpenAI GPT into a live voice call pipeline to enable immediate end‑to‑end testing.
- Design and implement a custom LLaMA 3 8B model, including architecture decisions, data generation, and training pipelines.
- Generate and curate 10,000+ synthetic NHS GP conversation examples for model fine‑tuning.
- Apply LoRA, supervised fine‑tuning (SFT), and reinforcement learning from human feedback (RLHF/DPO) to improve model performance.
- Deploy, monitor, and scale the fine‑tuned model on AWS SageMaker, ensuring reliability and compliance with healthcare data standards.
Requirements
- Strong proficiency in Python and deep learning frameworks, especially PyTorch.
- Hands‑on experience with large language models (e.g., LLaMA, GPT) and fine‑tuning techniques such as LoRA and RLHF.
- Proven ability to build and manage data pipelines for synthetic data creation and model training.
- Experience deploying ML models on cloud platforms, preferably AWS SageMaker.
- Understanding of NLP for voice applications and familiarity with healthcare data privacy regulations.