3. Data, Science & AI

Guides users through creating and configuring Claude Code subagents for task delegation, including tool selection and system prompts.

A programmable system for creating, splitting, and configuring AI agent teams.

Manages data ingestion, storage layout, and metadata integrity for HDF5 files in data pipelines.

Creates LLM-powered conversational AI assistants with integrated tools and data sources for database and API interactions.

Provides calibration, flagging, imaging, and mosaicking workflows for producing science-quality radio astronomy data products.

Calculates water quality scores using mathematical logic, enabling adjustments for new minerals and threshold changes with clear result explanations.

Retrieves MLflow experiment traces via CLI or Python API with filtering by ID, status, metadata, and execution time for debugging and analysis.

Provides safe deletion protocols for dictionary entries while maintaining index integrity and cross-reference consistency.

Automates quality assurance checks, metrics, and regression verification for data pipeline outputs, including image quality assessment.

Creates interactive dashboards for visualizing progress metrics and reporting data effectively.

Guides creation, improvement, and configuration of specialized AI agents, covering model selection, tool setup, and best practices.

Provides standards, templates, and validation tools for creating interoperable AI agent skills across different platforms.

Expertly designs and optimizes relational database architectures, covering schema planning, indexing, sharding, and multi-tenant isolation for performance and scalability.

Provides a structured framework for validating, cleansing, and monitoring data quality across datasets.

Integrates Olakai monitoring SDK to instrument existing AI agents and LLM calls for observability and telemetry.

Transcribes audio to text with timestamps and speaker identification for meetings, subtitles, and voice recordings.

Provides a comprehensive overview of data management practices, including processing, storage, and analysis techniques.

Detects unsafe training load spikes and applies conservative safety adjustments via emitted SafetyFlags.

Intelligently coordinates AI agents to execute complex tasks through strategic planning, phased breakdown, and quality verification.

Generates structured reflections from run telemetry and user notes, including confidence scores and safety checks for pain or abnormal heart rate to guide next steps.

Detects unsafe training load spikes (>20-30% week-over-week) and provides safety flags with conservative adjustment recommendations for model training.

Guides instrumenting Python and TypeScript code with MLflow Tracing for observability in LLM applications, agents, and frameworks like LangChain.

Provides structured guidelines for migrating data between systems while ensuring integrity and minimizing disruption.

Analyzes run telemetry to generate actionable insights, recovery recommendations, and session nudges for system optimization.