SlipstreamJobs tracks this role from the company's public career site. Apply directly on the employer's site.
Iambic Therapeutics is seeking a Machine Learning Scientist to join the Enchant team, working on post-training methods for large multimodal transformer models applied to drug discovery. The role focuses on researching and developing advanced post-training approaches including supervised fine-tuning, parameter-efficient fine-tuning (LoRA), reinforcement learning (RLHF, RLAIF, PPO, GRPO), and reward-based optimization for biomedical foundation models.
Key responsibilities include designing and evaluating post-training strategies for multimodal LLMs, creating reward functions and training objectives for reinforcement learning applications, building systematic experimentation and hyperparameter optimization workflows using tools like Optuna and Ray Tune, and developing inference optimization techniques for both high-throughput evaluation and interactive discovery workflows. You will design rigorous benchmarking frameworks to measure model quality across modalities and scientific use cases, collaborate with ML and software engineering teams to productionize models and evaluation systems, and partner with computational and medicinal chemists to ground model development in real drug discovery needs.
The role requires strong Python and PyTorch expertise, demonstrated experience training large-scale transformer models, and proficiency in modern ML infrastructure (Docker, CUDA, Kubernetes, Weights & Biases). Preferred qualifications include experience with multimodal architectures, training/inference optimization techniques (mixed precision, quantization, distributed strategies), familiarity with biomedical or chemical data domains, and distributed training at scale. You will communicate research results internally and at conferences, and write high-quality, well-tested, documented code.
Iambic is a clinical-stage biotech company founded in 2020, based in San Diego, combining AI expertise with drug discovery experience. The company has demonstrated rapid delivery of drug candidates to human clinical trials across multiple target classes. This is a remote position with optional on-site access to the Bristol office.
