3. Data, Science & AI

Diagnoses AI model errors in Pollinations services by analyzing logs, identifying error patterns, and pinpointing affected users.

Analyzes revenue, pack purchases, and tier spending patterns by querying Polar for payment history and Tinybird for usage data.

Provides a syntax guide for writing prompts in the Latitude platform, covering variables, conditionals, loops, chains, tools, and agents for effective LLM prompt engineering.

Provides end-to-end workflow guidance for creating, implementing, and testing goal-driven AI agents, orchestrating related skills for new projects.

Step-by-step guide for constructing goal-driven AI agents, including package setup, goal definition, node addition, and agent class finalization.

Provides goal-based testing for AI agents to verify goal achievement, debug failures, and iteratively improve agent performance.

Provides core concepts for building goal-driven AI agents, including architecture, node types, tool discovery, and workflow overview.

Provides best practices, patterns, and anti-patterns for designing goal-driven AI agents with hybrid workflows and pause/resume architecture.

Intelligently selects optimal chart type from 26 options and generates chart image from data specifications using Python.

Manages AI SDK model configurations, including package updates, missing model identification, new model integration with research, and documentation maintenance.

Monitors Beaker experiments until completion, providing real-time status updates for machine learning experiment tracking.

Automates specialist agent selection for user requests via AI analysis, eliminating explicit agent mentions.

Queries the memory system to retrieve relevant information from previous interactions, enhancing context-aware AI responses.

Searches Mathlib for lemmas matching type signature patterns, supporting formal verification and theorem proving in Lean.

Builds and executes LLM-powered pipelines for unstructured data analysis, document processing, and text-based ETL tasks.

Configures and optimizes analytics tracking systems including GA4, conversion tracking, and UTM parameters for data collection and measurement.

Provides problem-solving strategies for entropy calculations and applications in information theory, relevant to data science and AI tasks.

Provides mathematical strategies for solving boundary value problems in ordinary and partial differential equations, essential for scientific and engineering simulations.

Enables creation and configuration of custom sub-agents in Claude Code for tailored AI interactions via personalized prompts, tools, and models.

Provides problem-solving strategies for matrix operations and linear algebra computations, aiding in mathematical analysis and scientific applications.

Automates tracing of Claude Code conversations to Braintrust, capturing sessions, turns, and tool calls for AI observability.

Spawns multi-agent patterns using the Agentica framework for AI system coordination and collaboration.

Offers problem-solving strategies for second-order ordinary differential equations (ODEs) and partial differential equations (PDEs).

Provides problem-solving strategies for convex optimization problems, essential in machine learning and data science applications.