Associate Principal AI Scientist, AstraZeneca

Associate Principal AI Scientist (Drug Development)

Company: Pharmaceutical Technology and Development (PT&D) Department

Location: Durham, North Carolina, USA

About the Role

We are seeking two Associate Principal AI Scientists eager to revolutionize drug development processes using advanced AI technologies. Within the PT&D department, you will be instrumental in transforming molecules into groundbreaking medical treatments. Your role involves applying data science expertise to global pharmaceutical development projects, shaping our AI strategy, and driving the co-development of sophisticated Human-in-the-Loop (HITL) multi-agent systems.

Essential Skills and Experience:

• Education: Minimum Bachelor's degree in Computer Science, Data Science, Artificial Intelligence, Machine Learning, or related fields.
• Experience: At least 3 years of experience in Deep Learning and ML.
• Coding: Excellent coding skills in languages such as Python, R.
• Multi-Agent Systems: Hands-on industrial experience designing multi-agent patterns, digital twins, and agentic AI design patterns, including reinforcement learning.
• AI/ML Frameworks: Extensive industrial experience with AI and ML frameworks like TensorFlow, PyTorch.
• GenAI Orchestration: Hands-on experience with GenAI orchestration frameworks such as LangGraph, CrewAI.
• Reinforcement Learning: Hands-on experience with reinforcement learning libraries such as OpenAI Gym, Ray RLlib, or Stable Baselines.
• Applied ML Domains: Hands-on industrial experience with applied machine learning domains including deep learning, NLP, and GenAI.

Desirable Skills and Experience:

• Contributions to open-source projects (highlight merged GitHub PRs).
• Strong publication record in the field of AI.
• Experience designing multi-agent systems specifically within the pharmaceutical sector.
• Experience delivering ML projects in pharmaceutical development, chemical engineering, or chemistry.
• Experience with advanced ML domains: transfer learning, federated learning, few/zero-shot learning, meta-learning, explainable AI.