domains / research / ai-ml-skill-routing / ai-ml

ai-ml

Skill card

combined

The two copies are different routing documents for the same broad AI/ML domain. The first is a production-oriented workflow bundle that orchestrates named skills across LLM integration, RAG, agents, ML pipelines, observability, and security, and it includes explicit selection/completion constraints and checklists. The second is a research-oriented index of 27 AI/ML skills focused on papers, benchmarks, frameworks, and model training, with a simple select-and-read instruction. Neither copy is a reproduction of the other, and crowning one would hide a real alternative: production workflow orchestration versus research skill discovery.

Repositories

Other copies

Combined skill

---
name: ai-ml
---

# AI/ML Skill Routing

Select the routing mode that matches the user's request, then follow the corresponding steps.

## Production AI/ML Workflow

Use this mode when the user is building LLM applications, RAG systems, AI agents, ML pipelines, AI-powered features, or AI observability.

1. Select only the phases needed for the requested product. This workflow does not authorize delegation, hosted calls, deployment, background jobs, or external messages.
2. Read the actual selected skill and available tool schema.
3. Define task-level quality, retrieval/source checks, data permissions, latency/currency limits, and failure recovery before implementation.
4. Verify each used capability in the target environment. Unchecked checklist items are not completed because an instruction exists.

### Phase 1: AI Application Design
- Skills: ai-product, ai-engineer, ai-agents-architect, llm-app-patterns
- Actions: define AI use cases, choose models, design system architecture, plan data flows, define success metrics.

### Phase 2: LLM Integration
- Skills: llm-application-dev-ai-assistant, llm-application-dev-langchain-agent, llm-application-dev-prompt-optimize, gemini-api-dev
- Actions: select LLM provider, set up API access, implement prompt templates, configure model parameters, add streaming support, implement error handling.

### Phase 3: RAG Implementation
- Skills: rag-engineer, rag-implementation, embedding-strategies, vector-database-engineer, similarity-search-patterns, hybrid-search-implementation
- Actions: design data pipeline, choose embedding model, set up vector database, implement chunking strategy, configure retrieval, add reranking, implement caching.

### Phase 4: AI Agent Development
- Skills: autonomous-agents, autonomous-agent-patterns, crewai, langgraph, multi-agent-patterns, computer-use-agents
- Actions: design agent architecture, define agent roles, implement tool integration, set up memory systems, configure orchestration, add human-in-the-loop.

### Phase 5: ML Pipeline Development
- Skills: ml-engineer, mlops-engineer, machine-learning-ops-ml-pipeline, ml-pipeline-workflow, data-engineer
- Actions: design ML pipeline, set up data processing, implement model training, configure evaluation, set up model registry, deploy models.

### Phase 6: AI Observability
- Skills: langfuse, manifest, evaluation, llm-evaluation
- Actions: set up tracing, configure logging, implement evaluation, monitor performance, track costs, set up alerts.

### Phase 7: AI Security
- Skills: prompt-engineering, security-scanning-security-sast
- Actions: implement input validation, add output filtering, configure rate limiting, set up access controls, monitor for abuse, implement audit logging.

### Checklists
- LLM Integration: API keys secured, rate limiting configured, error handling implemented, streaming enabled, token usage tracked.
- RAG System: data pipeline working, embeddings generated, vector search optimized, retrieval accuracy tested, caching implemented.
- AI Agents: agent roles defined, tools integrated, memory working, orchestration tested, error handling robust.
- Observability: tracing enabled, metrics collected, evaluation running, alerts configured, dashboards created.
- Quality Gates: all AI features tested, performance benchmarks met, security measures in place, observability configured, documentation complete.

## Research AI/ML Skill Index

Use this mode when the user needs ML experiments, model training, deep learning, NLP, computer vision, paper reproduction, benchmarks, or AI research workflows.

Select the skill matching the user's need, then read its SKILL.md.

| Skill | Description |
|-------|-------------|
| ai-agent-papers-guide | Curated 2024-2026 AI agent research papers collection |
| ai-model-benchmarking | Benchmark AI models across 60+ academic evaluation suites and metrics |
| anomaly-detection-papers-guide | Industrial anomaly detection methods and benchmark papers |
| autonomous-agents-papers-guide | Daily-updated collection of autonomous AI agent papers |
| computer-vision-guide | Apply computer vision research methods, models, and evaluation tools |
| deep-learning-papers-guide | Annotated deep learning paper implementations with code walkthroughs |
| dl-transformer-finetune | Build transformer fine-tuning plans for classification and generation |
| domain-adaptation-papers-guide | Comprehensive collection of domain adaptation research papers |
| generative-ai-guide | Curated guide to generative AI covering LLMs and diffusion models |
| graph-learning-papers-guide | Conference papers on graph neural networks and graph learning |
| huggingface-api | Search and discover ML models, datasets, and Spaces on Hugging Face |
| huggingface-inference-guide | Run NLP and CV model inference via Hugging Face free-tier API |
| keras-deep-learning | Build and debug deep learning models with Keras and TensorFlow backend |
| kolmogorov-arnold-networks-guide | Papers and tutorials on KAN learnable activation networks |
| llm-evaluation-guide | Evaluate and benchmark large language models for research applications |
| llm-from-scratch-guide | Build a ChatGPT-like LLM from scratch using PyTorch step by step |
| ml-pipeline-guide | Build and deploy reproducible production ML pipelines for research |
| nlp-toolkit-guide | NLP analysis with perplexity scoring, burstiness, and entropy metrics |
| npcpy-research-guide | All-in-one Python library for NLP, agents, and knowledge graphs |
| prompt-engineering-research | Systematic prompt engineering methods for AI-assisted academic research workflows |
| pytorch-guide | Avoid common PyTorch mistakes and apply robust training patterns |
| pytorch-lightning-guide | PyTorch Lightning framework for scalable model training and research |
| reinforcement-learning-guide | Reinforcement learning fundamentals, algorithms, and research |
| responsible-ai-guide | Resources for trustworthy, fair, and ethical AI research |
| tensorflow-guide | TensorFlow best practices for tf.function, GPU memory, and deployment |
| transformer-architecture-guide | Guide to Transformer architectures for NLP and computer vision |
| vmas-simulator-guide | Vectorized multi-agent reinforcement learning simulator |

## Conflicts
- sickn33/agentic-awesome-skills defines a multi-phase production workflow with explicit constraints on delegation, hosted calls, deployment, background jobs, external messages, and checklist completion. brycewang-stanford/Auto-Empirical-Research-Skills provides a flat research skill index with a simple select-and-read instruction and no such constraints.