domain-ml

What problem does it solve?

This Skill provides domain-aware guidance to transform ML challenges in Rust into scalable, well-architected solutions by tracing constraints into concrete design and implementation patterns.

Core Features & Use Cases

• Domain-guided design: translate constraints such as memory, GPU utilization, and reproducibility into architecture decisions and safe Rust code patterns.
• Inference server patterns: lazy initialization, singleton model loading, batched inference, and streaming data handling for low-latency Rust services.
• Data processing integration: leverage crates like ndarray and polars for tensor and data-frame operations, and integrate ONNX/tch-rs models for inference.

Quick Start

Prompt the AI with: Design a Rust-based ML inference service that loads an ONNX model with tract, uses batched input, and keeps a lazy singleton in memory.