Build agentic LLM systems that acquire, clean, validate, curate, and document large-scale biomedical datasets for multimodal model training. Develop reliable agent architectures, tooling, evaluation harnesses, quality-control workflows, and reproducible data pipelines. Collaborate with ML scientists, monitor distributed production pipelines, diagnose failures, and implement secure sandboxed execution with restricted access and audit logging.

Research and develop post-training methods for large multimodal transformer models used in drug discovery. Responsibilities include supervised and parameter-efficient fine-tuning, reinforcement learning, reward design, experimentation, hyperparameter optimization, inference optimization, benchmarking, and productionization. The role involves collaborating with machine learning engineers, software developers, computational chemists, medicinal chemists, and biologists while writing robust, tested, and documented research and engineering code.

Research and develop post-training methods for large multimodal transformer models used in drug discovery. Responsibilities include reinforcement learning, fine-tuning, reward design, experimentation, inference optimization, benchmarking, and productionizing ML systems. The role collaborates with software engineers, machine learning researchers, computational chemists, medicinal chemists, and biologists to align model objectives with therapeutic applications. Candidates will also write, test, document, and communicate high-quality research and engineering work.