3. Data, Science & AI

Manages DeFi registry data including tokens, chains, DEXes, and protocols for seamless expansion and updates.

Analyzes model checkpoint vulnerability to reward hacking through prefill sensitivity, token trajectories, and logprob vs token-count comparison.

Generates data visualizations and charts using multiple charting libraries for effective data representation.

Quick reference guide for common SQL patterns, including CTEs, window functions, and indexing strategies for efficient database querying.

Generates PACKMOL input files for solvating proteins and building initial configurations in molecular dynamics simulations.

Provides a systematic framework to select between AI coding tools (Claude Code and Gemini CLI) based on task requirements, codebase size, and complexity, ensuring optimal resource allocation.

Proactively retrieves prior conversation context, decisions, and discussions to inform new interactions, improving AI response coherence.

Intelligently processes large document corpora using knowledge graphs and multi-hop reasoning to enable complex cross-document queries and persistent knowledge retrieval.

Converts PDF pages to semantic HTML using AI, processing PNG, text, and ASCII inputs for precise structural reconstruction.

Optimizes prompts for LLMs and AI systems to enhance performance, accuracy, and interaction quality in AI feature development.

Automates web scraping, structured data extraction from JavaScript-heavy pages, and schema-based data pipeline generation without LLMs.

Validates data against schemas, business rules, and data quality standards to ensure integrity and compliance.

Expert assistance in writing, optimizing, and understanding Datalog queries for Logseq's graph database, covering advanced patterns and Datascript engine techniques.

Coordinates multiple AI agents in real-time using low-latency tmux or file-based messaging for synchronous and asynchronous workflows.

Provides expert domain knowledge in Physical AI, ROS 2, and Humanoid Robotics for advanced robotic system development.

Disables continuous learning mode, preventing AI systems from adapting to new interactions or data inputs.

Analyzes Excel spreadsheets, creates pivot tables, generates charts, and performs data analysis for tabular data.

Transforms data arrays using standard operations including filtering, mapping, and aggregation for data preparation.

Executes PostgreSQL database queries and administrative commands through the command-line interface.

Optimizes vector index performance for latency, recall, and memory through HNSW tuning and quantization strategies.

Provides enterprise charting components using G2Plot for data visualization in Angular, specifically for construction site progress tracking systems.

Enables creation and training of AI learning plugins using AgentDB's reinforcement learning algorithms for self-learning agents and behavior optimization.

Enables building, training, evaluating, and deploying machine learning models and pipelines for production systems.

Analyzes images using Zhipu AI's GLM-4V model for content recognition and visual question answering.