3. Data, Science & AI

Provides workflow patterns and common pitfalls for LangGraph, guiding RAG implementation in AI systems.

Provides workflow patterns for orchestrating multi-agent systems, with implementation guidance for RAG (Retrieval-Augmented Generation) in AI applications.

Guides users in configuring ADK's visual builder for no-code agent development and advanced reasoning features like chain of thought.

Guides implementation of OpenAI Agents using RAG patterns, addressing common pitfalls and workflow best practices.

Guides on workflow patterns and pitfalls for CrewAI framework usage, with implementation direction to RAG for AI solutions.

Generates synthetic multi-turn conversation data for AI model fine-tuning, requiring prior design artifacts from 'finetune-design'.

Prepares for LLM fine-tuning by defining evaluation criteria and designing conversation data generation for multi-turn dialogues.

Evaluates base models for domain-specific tasks to determine if fine-tuning is required, ensuring quality thresholds are met.

Trains a fine-tuned model on filtered data, evaluates performance against a base model, and converts to GGUF format.

Guides on LangChain workflow patterns and common pitfalls for RAG implementation.

Adds a new Evidence.dev dashboard page featuring data charts and tables for visualization.

Initializes Kaggle/atmaCup competition projects with pre-configured data science environments and project structure.

Provides a web-based dashboard for real-time telemetry visualization, live configuration, field overlay drawing, and camera streaming for FTC robotics teams.

Scaffolds a new Delta Live Tables (DLT) pipeline for data ingestion into MotherDuck, enabling streamlined data pipeline setup.

Extracts and analyzes text and form data from PDF documents.

Assists in crafting and refining AI prompts, instructions, and system messages for enhanced model performance and clarity.

Enables Neo4j agents to self-report required skill updates for continuous capability improvement in AI systems.

Validates and filters data from sub-agents, returning only clean, reliable data for downstream applications.

Executes user prompts through OpenAI or Gemini AI models using a command-line interface for efficient AI interactions.

Provides a reference guide for selecting the correct Perplexity tool before initiating any agent, optimizing AI workflow efficiency.

Queries earnings call transcripts, Q&A exchanges, and prepared remarks stored in Neo4j for business data analysis and insights.

Reference tool for Alpha Vantage API data endpoints, enabling exploration of available financial data.

Retrieves SEC filing data including 8-K, 10-K, 10-Q reports, sections, exhibits, and financial statements from Neo4j.

Accesses consensus earnings estimates and historical earnings data through the Alpha Vantage API for financial analysis.