3. Data, Science & AI

Provides runtime interface for Amazon Bedrock model inference, supporting Claude, Titan, and third-party models with streaming, token counting, and guardrails.

Proposes and updates system rules based on detected patterns and explicit user requests to enhance AI response accuracy.

Integrates Nano-Banana AI service for content and image generation workflows using AI-based models.

Expert guidance for building AI agents with Microsoft Agent Framework, covering single agents, multi-agent workflows, and LLM tool integration using Python and .NET.

Updates and maintains AI memory bank files including context, progress, and decision logs for seamless system state management.

Leverages Gemini and Qwen CLI tools for AI-driven image analysis, including description, text extraction, and visual summarization with fallback mechanisms.

Preprocesses library content (PDFs, EPUBs, academic papers) by cleaning, optimizing for speed reading, and fixing extracted text to enhance readability and efficiency.

Manages, shares, and syncs AI agent skills across platforms with a single installation.

Builds knowledge models from scratch by defining core classes, relationships, and taxonomies for AI and data systems.

Provides global weather data including 16-day forecasts, historical climate data, air quality, and marine conditions through the Open-Meteo API.

Customizes Amazon Bedrock models via fine-tuning, distillation, and continued pre-training to enhance domain-specific accuracy and performance.

Enables AI agents to maintain episodic, procedural, semantic, and short-term memory across sessions for adaptive learning and context preservation.

Integrates xAI Grok sentiment with financial APIs (FinnHub, Twelve Data, Alpha Vantage, FMP) for combined market and sentiment analysis.

Designs AI architectures, crafts effective prompts, and builds RAG systems and LangChain applications for LLM-based solutions.

Manages persistent memory for autonomous coding systems, enabling storage, retrieval, and integration of knowledge and episodic memory.

Supports A/B testing and experiment tracking with hypothesis validation and statistical analysis for data-driven decision making.

Automates data cleaning, statistical analysis, and visualization using Python's Pandas and Matplotlib libraries for structured data workflows.

Guides users through selecting model export formats and initiating the export process for AI/ML models.

Enables AI-powered image analysis, visual content description, and multimodal interaction through vision-language models.

Automates context summarization for AI sessions to manage token limits, preserve critical information, and maintain efficiency in long conversations.

Systematically diagnoses and resolves TensorFlow/Keras machine learning issues including tensor mismatches, GPU errors, and model loading problems.

Provides expert guidance on designing data platforms, optimizing pipelines, and establishing engineering standards for data systems.

Orchestrates multi-agent systems via Amazon Bedrock's AgentCore with A2A protocol, enabling supervisor-worker patterns and hierarchical delegation for complex AI workflows.

Converts Product Requirements Documents (PRDs) into the structured prd.json format required by the Ralph autonomous agent system.