3. Data, Science & AI

View and analyze Hawk AI evaluation results, including status, samples, transcripts, and agent behavior from completed runs.

Manages AI command-line tools including version checks, updates, and installation instructions for Claude Code, Gemini CLI, and Codex CLI.

Designs production-ready machine learning systems covering data pipelines, feature stores, model training, deployment, and monitoring.

Designs database schemas using dimensional modeling, star schema, and entity-relationship techniques for efficient data storage and retrieval.

Optimizes vector database selection, embedding storage, and approximate nearest neighbor search for semantic and similarity-based AI applications.

Guides the design of data platforms by selecting appropriate architectures including data lakes, lakehouses, warehouses, and implementing data mesh patterns.

Generates meta-prompts that create other prompts for AI applications, enabling prompt generator development and template scaffolding for advanced AI workflows.

Optimizes AI prompts using Anthropic's best practices with a 4-step workflow, decision trees, and chain of thought refinement for enhanced model accuracy.

Creates YAML-based expertise schemas for domain-specific AI agents, enabling structured mental models and Act-Learn-Reuse knowledge workflows.

Guides strategic codebase exploration using Gemini's large context window, covering token thresholds, model routing, and decisions between Gemini and Claude.

Optimizes Gemini CLI token usage via caching, batch queries, model selection, and cost tracking for efficient bulk AI operations.

Designs and implements real-time data processing systems using stream processing frameworks like Kafka and Flink for event-driven architectures.

Guides creation of AI Developer Workflows (ADWs) integrating deterministic code with non-deterministic agents for automated pipelines and PITER framework implementation.

Audits codebases for 12 leverage points in agentic coding, identifying gaps and prioritizing recommendations for agent optimization.

Extracts and categorizes requirements from PDFs, Word documents, transcripts, and web content into pre-canonical format.

Creates reusable meta-prompt templates to encode engineering workflows for scalable AI-driven planning and team best practices.

Designs memory hierarchies for efficient context management in AI agents, enabling progressive loading and optimal context utilization.

Generates meta prompts, agents, and skills to automate creation of agentic components, enabling scalable AI system development via self-referential patterns.

Provides expert guidance for integrating Model Context Protocol (MCP) servers with Gemini CLI, covering configuration, connection management, and tool permissions.

Designs reliable agentic workflows using Request-Validate-Resolve prompt structure to ensure self-validation and feedback loops in AI agent operations.

Builds monitoring interfaces for multi-agent AI systems, enabling metric tracking, logging, and debugging of parallel agent workflows.

Selects optimal AI models (Haiku, Sonnet, Opus) for agent tasks by balancing cost, quality, and task complexity.

Designs media asset management systems including storage patterns, metadata schemas, and headless CMS APIs for digital media workflows.

Designs RAG pipelines with optimal chunking, embedding models, and retrieval techniques for enhanced LLM knowledge grounding.