3. Data, Science & AI

Manages Hugging Face Hub datasets with repo initialization, config definition, streaming updates, and SQL-based transformations for AI workflows.

Publishes and manages AI research papers on Hugging Face Hub, enabling model/dataset linking and professional markdown article generation.

Tracks and visualizes machine learning training experiments with real-time dashboards, HF Space integration, and CLI for metric analysis.

Trains and fine-tunes language models via Hugging Face Jobs with TRL methods (SFT, DPO, reward modeling) and GGUF conversion for deployment.

Manages and adds model evaluation results to Hugging Face model cards, supporting README extraction, API imports, and vLLM/lighteval evaluations.

Executes Hugging Face Hub operations via CLI, enabling model/dataset downloads, repository management, and cloud compute job execution.

Enables invocation of AWS Bedrock foundation models, creation of embeddings, and implementation of RAG patterns for generative AI applications.

Tests network connectivity to AI providers including Gemini, Codex, and Claude for service availability.

Delegates user requests to external AI providers (Gemini, Codex, OpenCode, Droid) without further processing, ending the current turn immediately.

View the latest responses from AI providers including Gemini, Codex, and Claude for quick reference.

Delegates user queries to external AI providers (Gemini, Codex, etc.) for immediate async responses, without handling provider-related questions.

Tests connectivity to major AI service providers including Gemini, Claude, and others to verify API availability and responsiveness.

Comprehensive guide for implementing AI model integrations using CloudBase's JS/Node SDK and WeChat Mini Program, supporting text/image generation with streaming.

Provides a Web SDK for CloudBase document database operations, enabling CRUD, complex queries, pagination, and geolocation data management.

Enables WeChat MiniPrograms to perform database operations including CRUD, complex queries, and geolocation features via CloudBase SDK.

Provides advanced capabilities for designing complex data models, complementing simpler SQL-based table creation tools.

Enables AI text generation and streaming in WeChat Mini Programs using built-in models like Hunyuan and DeepSeek.

Documentation for safe SQL execution and security rule management on CloudBase Relational Database to prevent data corruption during CRUD operations.

Builds neuro-symbolic LLM applications using the Synalinks framework, supporting structured outputs, RAG, and Keras-like workflows for AI pipelines.

AI-powered classifier that analyzes text prompts to accurately determine domain (portrait, art, design, product, video) using natural language processing.

Analyzes AI prompts to provide insights, compare variations, recommend similar prompts, and offer element library statistics for improved model interaction.

Automatically extracts reusable elements from domain-agnostic prompts and accumulates knowledge for continuous AI learning.

Automates extraction and structuring of modular components from AI art prompts for reuse in image generation workflows.

Reverses engineered from high-quality prompts to extract 'how-to' methodology, making AI prompt design transparent and actionable.