Show HN: Easy to use Cluster-Compute software (docs.burla.dev)

Burla is an open-source cluster-compute platform that promises to make scaling Python trivial with a single API: remote_parallel_map. With that one call you can dispatch a Python function across hundreds to thousands of machines (the project advertises up to 10,000 CPUs and GPU support), use custom Docker images, and have a cloud storage bucket automatically mounted as a network filesystem. Burla forwards prints and exceptions back to your local terminal so remote execution feels like local coding, and it exposes per-call resource controls (func_cpu, func_ram) to assign heterogeneous hardware to different tasks. Nested remote_parallel_map calls let you build multi-stage pipelines (e.g., per-record work across many files, then reduce on a large machine) without special orchestration syntax. For the AI/ML community this lowers the barrier to parallelizing data processing, hyperparameter sweeps, and distributed model-training experiments—developers can scale experiments quickly without weeks of ops onboarding. Technically notable points are the single-call UX, automatic cloud-storage mounts, per-function resource sizing, and support for GPUs and custom containers. Practical implications include faster iteration and simpler pipeline composition, but teams will still need to consider data locality, network transfer costs, security of executing arbitrary code across many hosts, and resource accounting when moving from prototypes to production.