Machine learning

World-class prompt engineering skill for LLM optimization, prompt patterns, structured outputs, and AI product development.

Enforces structured, highly documented storage for code and data projects.

Use this skill for processing and analyzing large tabular datasets (billions of rows) that exceed available RAM.

Designs complex generic types, refactors `any` types to strict alternatives, creates type guards and utility types, and resolves TypeScript compiler errors.

Use for "interrogate", "adversarial review", "multi-model review", "challenge this", "stress test this code", "find blind spots", or "tear this apart".

Apply source-aware mechanical-engineering judgment to research, analysis, coding, writing, teaching, research identity, and release work.

Best practices for scikit-learn machine learning, model development, evaluation, and deployment in Python

Investigate outliers, rare events, spikes, and suspicious records in datasets.

Evaluate trained machine learning models with the right metrics and comparison logic.

Explain trained machine learning models through feature attribution, local explanations, and behavior summaries.

Design and implement repeatable preprocessing pipelines for cleaning, encoding, transforming, and validating ML input data.

Deep learning framework (PyTorch Lightning).

Split datasets into training, validation, and test partitions with the right stratification and temporal rules.

Graph-based drug discovery toolkit.

Hierarchy → 精簡結構化文字匯出工具（HierarchyTextExporter）。當需要：(1) 讓 LLM 用最少 token 讀懂 Unity GameObject 階層與 Component serialized 欄位 (2) 匯出 prefab 或 scene 子樹成精簡文字（uloop…

Use when building self-correcting retrieval systems for AI agents.

Use when optimizing token usage, KV cache efficiency, or context window management for LLM agents.

Use when needing global or relational understanding of codebases or knowledge corpora.

Use when validating, benchmarking, or monitoring LLM application performance.

Use when the user wants to optimize configurable system parameters against a measurable scalar objective, especially for model training, inference…

Bundle the canonical memory set into one structured Markdown document for external-LLM review, audit, or dashboard download.

Inspect and manage Hugging Face Hub models, datasets, Spaces, repositories, cache, authentication, and downloads with the current hf CLI.

Explore, validate, filter, paginate, export, and prepare Hugging Face datasets through the Dataset Viewer API and Hub tooling.

Plan, launch, inspect, monitor, and cancel authenticated Hugging Face Jobs for CPU, GPU, TPU, batch, inference, and data workloads.