3. Data, Science & AI

Guides implementation of a 3D memory system using static knowledge graphs, timeline logs, and Milvus vector retrieval for efficient knowledge organization.

Guides users in crafting effective prompts for SiftCoder to optimize AI-powered code generation outcomes.

Normalizes PostgreSQL database schemas and migrations, including pgvector setup and Render compatibility checks.

Indexes reference documents for RAG systems by extracting text from PDFs and building a searchable corpus for retrieval.

Provides expert guidance for database design, schema management, and data persistence strategies.

Builds document ingestion pipelines with parsing, embedding generation, and topic categorization for efficient semantic document management.

Builds autonomous AI agents using Vercel AI SDK with tool integration, multi-step workflows, and loop control for LLM applications.

Manages persistent context, decisions, and project learnings across sessions to maintain continuity in AI agent workflows.

Queries the project's RAG corpus to retrieve stored decisions, documentation, and patterns for reference.

Identifies available access methods for data sources including APIs, databases, and file systems.

Generates and configures Model Context Protocol (MCP) servers to extend AI agent capabilities with context-aware model integration.

Looks up SNP genotypes from SelfDecode using JWT authentication, expanding genetic analysis beyond 23andMe raw data coverage.

Automates research and updating of lightweight LLM models to latest versions via web search, ensuring codebase uses current implementations.

Analyzes natural language specifications to extract structured, testable features for software development and testing.

Provides patterns for building AI features with Vercel AI SDK v5, including migration from v4.

Defines structured communication protocols between AI agents and humans to optimize development workflows and ensure efficient interaction.

Governance for LLM prompts including versioning, documentation, and separation from code for AI application development.

Backend system for bioinformatics data tagging and classification, triggered by terms including DNA, classification, and search optimization.

Manages Supabase database access, security policies, and schema modifications for the Triatu application.

Manages AI model routing strategies for hybrid architectures, selecting between models like Gemini, MiniMax, and Dify based on user queries.

Searches the web and data registries to locate and retrieve data sources for analysis and processing.

Optimizes generative AI search engines (ChatGPT, Claude, Perplexity) for enhanced information retrieval performance and relevance.

Analyzes images, music, and video files using Gemini API via local paths or URLs for AI-driven insights.

Creates new AI agent skills adhering to the Agent Skills specification, enabling structured development of custom AI capabilities and patterns.