3. Data, Science & AI

Provides general utilities for data processing and manipulation tasks.

Provides utilities for common data processing, transformation, and cleaning operations.

Routes users to the correct PyTorch specialist skill based on symptoms and problem type.

Selects optimal neural network architectures (CNNs, Transformers, RNNs, GANs, GNNs) based on data modality and constraints.

Specializes in LLM implementation techniques such as prompt engineering, fine-tuning, RAG, and safety evaluation for AI models.

Directs users to guides covering ML model deployment, optimization, MLOps, and monitoring best practices in production environments.

Routes AI/ML tasks to correct Yzmir pack for frameworks, training, RL, LLMs, and production deployment.

Routes users to appropriate deep reinforcement learning (RL) approaches based on problem type and algorithm family.

Enables dynamic neural network topologies that grow, prune, or adapt during training for continual learning and modular composition.

Defines standardized requirements for SME agents including fact-finding, output contracts, confidence assessment, and advice qualification.

Provides a router for simulation math tasks including ODEs, state-space, stability, control, numerics, chaos, and stochastic processes.

Creates Gazebo simulation worlds with SDF, configuring ground planes, physics, and lighting for robotics applications.

Retrieves web content via fetch API and converts it to Markdown for real-time data access and analysis.

Fetches real-time and historical weather data including forecasts and air quality from professional APIs for any city.

Provides a read-only workflow for processing large data corpora using JSONL REPL server commands, minimizing context bloat and generating cited architecture and audit reports.

Constructs queryable capability graphs from manifests through AI-driven relationship inference among skills, resources, and capabilities.

Builds and optimizes data pipelines for ETL processes, ensuring scalable data infrastructure for analytics and AI applications.

Manages design, construction, and maintenance of knowledge bases combining RAG and graph systems for intelligent information retrieval.

Provides data-driven insights for product decisions by tracking user metrics, analyzing behavior, and optimizing key performance indicators across acquisition, engagement, retention, and revenue.

Builds RAG systems for knowledge-intensive applications, document search, and Q&A using embedding strategies and vector stores.

Builds voice interfaces using speech recognition and text-to-speech technologies, enabling natural user interactions through voice.

Designs and orchestrates multi-agent AI systems for complex tasks requiring specialization, parallel processing, and collaborative problem-solving.

Creates interactive charts, dashboards, and visualizations using modern data libraries for data analysis and presentation.

Dynamically spawns subagents within an AI agent framework using the runAgent execution method.