Staff Research Engineer, Scientific Computing and ML/Physics Infrastructure

Your Impact at LILA Lila Sciences is seeking a Research Engineer, Scientific Computing and ML/Physics Infrastructure to help turn promising research tools into robust, scalable systems. This role bridges research and production: you will work with scientists and ML researchers who can prototype useful tools, then help make those tools efficient, distributed, fault tolerant, and usable across Lila's compute environments. The Molecular Intelligence team is building ML and physics-based infrastructure for drug discovery, including biophysics workflows, computational chemistry tools, cofolding models, low-data learning systems, simulation workflows, and agent-usable scientific pipelines. We need an engineer who can improve code quality, architecture, GPU efficiency, cluster portability, and operational reliability without slowing down research velocity. What You'll Be Building Take research tools, prototypes, and scientific workflows developed by scientists or academic-style researchers and make them scalable, efficient, and maintainable. Collaborate directly with computational biophysics, computational chemistry, and machine learning scientists to turn research workflows into scalable agent-usable systems. Build and support ML and physics infrastructure for model training, molecular simulation, data processing, and agent-executed scientific workflows. Ensure workflows run reliably across multiple clusters and compute environments. Improve GPU utilization, distributed execution, throughput, fault tolerance, and reproducibility for ML and scientific workloads. Architect larger-scale systems around research code, including job orchestration, retry behavior, monitoring, artifact handling, and workflow traceability. Optimize ML, physics, and pipeline code for performance and scalability. Maintain development and execution environments across local, cloud, and GPU-based systems. Package scientific tools into reusable services, workflows, or APIs that can be used by researche