Research Engineer (LLM Performance)

Location: London, United Kingdom; hybrid working with attendance in the office 3 days per week

Company

hirify.global develops AI models and computational systems for drug discovery, building on and beyond the AlphaFold system to advance digital biology and medicine.

What you will do

• Implement and optimize LLM post-training methods at scale on frontier models.
• Collaborate with scientists and engineers to translate research methods into production-ready systems.
• Evaluate and deploy frameworks for supervised fine-tuning, reinforcement learning, and LLM evaluation.
• Diagnose and resolve performance bottlenecks and communication overhead in distributed training and inference systems.
• Apply low-precision methods to balance model performance and accuracy in real-world drug-design programs.

Requirements

• Significant experience with large-scale distributed training of LLMs.
• Experience with a deep learning framework such as JAX or PyTorch.
• Knowledge of parallelism strategies and collective communication libraries such as NCCL.
• Good understanding of GPU architectures and performance concepts.
• Excellent collaboration skills.
• Ability to work in the London office 3 days per week under the hybrid working model.

Nice to have

• Experience with general LLM serving stacks.
• Knowledge of XLA, Triton, Pallas, CUDA, or similar accelerator DSLs and compilers.
• Experience optimizing ML accuracy with low-precision formats.
• Experience building, deploying, and maintaining production systems on GCP.
• Interest in chemistry and biology.

Culture & Benefits

• Interdisciplinary collaboration across science and engineering.
• A creative, iterative environment focused on rigorous scientific work.
• Shared values centered on thoughtfulness, bravery, determination, and collaboration.
• Equal employment opportunities and support for workplace accommodations.
• Hybrid working with regular in-person collaboration.