SWE - Distributed

The role

You will architect and build distributed compute infrastructure for ML data processing, training, and fine-tuning. You will optimize cluster observability, resource utilization, and cost-efficient compute solutions.

What you will do

• Design, implement, and optimize distributed compute infrastructure for ML data processing, training, and fine-tuning.
• Improve cluster observability, scheduling, and resource utilization (CPU/GPU/TPU).
• Research and implement cost-efficient compute solutions.
• Develop tools for monitoring, debugging, and performance tuning of large-scale ML workloads.
• Collaborate with ML engineers to accelerate training pipelines and reduce bottlenecks.
• Apply emerging technologies in distributed computing strategically.

Required qualifications

• Experience in building or working with distributed computing frameworks (e.g., Ray, Dask, Celery)
• Good grasp of parallel computing, job scheduling, and resource management
• Comfortable identifying and resolving performance issues in distributed systems (profiling, bottlenecks, network overhead)
• Implemented solutions using cloud compute platforms (AWS, GCP, Azure) and cluster orchestration (Kubernetes, Slurm)
• Familiar with popular ML frameworks (PyTorch, TensorFlow, or JAX)
• Familiar with MLOps best practices such as model deployment and GPU performance monitoring

About the company

Achira is building foundation models to solve problems in simulation for drug discovery. The company works on ML infrastructure at frontier scale.