🧠 MachineLearning & DataScience SKILLS

This repository contains a curated selection of Machine Learning, Data Science, MLOps, Agent Orchestration, and Generative AI skills for the Antigravity agentic programming assistant.

It is designed to serve as an atomic and self-contained workspace that optimizes the Token Budget during the development of purely data-driven projects.

---

📂 Repository Structure

The repository includes elements organized as follows:

• ml_theory_and_model_selection.md: Centralized theoretical guide on ML taxonomy, Deep Learning, RAG, and metric selection.

🤖 1. Machine Learning & Traditional AI

• scikit-learn: Predictive modeling and classical ML (regression, classification, random forests, SVM).
• ml-engineer: Deep Learning systems in production (PyTorch 2.x, TensorFlow, distributed training).
• data-scientist: Exploratory Data Analysis (EDA), applied statistics, and hypothesis testing.
• statsmodels: Rigorous statistical modeling in Python.
• vector-database-engineer: Vector indices and semantic similarity.
• rag-implementation: End-to-end RAG pipelines and ingestion.
• rag-engineer: Advanced RAG architecture patterns.

🧠 2. AI Engineering, LLMs & Autonomous Agents [NEW]

• ai-engineer: Building production-ready LLM applications, advanced RAG systems, and multimodal AI.
• multi-agent-architect: Design and optimization of production multi-agent systems with LangGraph and LangChain.
• agent-orchestrator: Orchestration of multi-skill workflows, capabilities, and technical agent coordination.
• prompt-engineering-patterns: Advanced prompt engineering patterns (hallucination minimization, output structuring).
• advanced-evaluation / agent-evaluation: LLM-as-a-judge metrics, bias mitigation, and agent benchmarking.
• deep-research: Deep autonomous research tasks (planning, information synthesis, and reporting).

📊 3. Data Processing and Visualization

• pandas-expert: Advanced data manipulation, vectorization, and memory optimization.
• polars: Fast data manipulation with Polars DataFrames.
• matplotlib: Generation of static plots and learning curves.
• plotly: Dynamic and interactive visualizations.
• dbt-transformation-patterns: Modular data modeling and engineering with dbt.
• database-design: Relational and storage database schema design.
• postgresql-optimization: Optimization of complex SQL queries and indexing.

⚙️ 4. MLOps, DevOps & Infrastructure

• ml-pipeline-workflow: End-to-end MLOps automation pipelines.
• docker-expert: Reproducible containers for models and runners.
• gitops-workflow: Continuous deployment of inference microservices.
• observability-engineer: Observability and data drift detection.
• grafana-dashboards: Infrastructure and health monitoring.
• cost-optimization: Cost optimization for storage and GPU instances.

🛠️ 5. Systems Architecture & Methodology

• senior-architect & backend-architect: Structuring of data and service layers.
• architecture-decision-records: Design trade-off logs (ADRs).
• mermaid-expert: Flowcharts and architecture diagrams in markdown.
• code-reviewer: Analytical code quality auditing.
• systematic-debugging: Step-by-step scientific debugging of complex bugs.
• planning-with-files: Structured management of experimentation milestones.
• satori & explain-like-socrates: Cognitive reasoning frameworks.

🧪 6. Testing & Quality

• tdd-workflow: Test-Driven Development (TDD) for decoupled code.
• python-testing-patterns: Software testing in Python using pytest.

---

🚀 How to use this Repository

You can copy the skills from this repository to your local Antigravity configuration folder: C:\Users\<your-username>\.gemini\config\skills

And add them to your $whitelist file in your filtering script to keep them always active during your Data Science and AI sessions.