3. Data, Science & AI

Specializes in debugging and implementing Airtable synchronization logic, including duplicate prevention, cache management, and change detection for reliable data integration.

Builds real-time voice AI agent backends in Python using LiveKit's Agents SDK, supporting STT, LLM, TTS, and multi-agent workflows.

Optimizes prompts for LLMs and AI systems to enhance model performance, agent behavior, and interaction quality.

Standardizes tool set definitions for Claude Code agents, ensuring consistent tool access across similar agent types.

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

Executes Python scripts for data analysis, statistical testing, visualization, and predictive modeling with reproducible workflows.

Comprehensive spreadsheet tool for data analysis, formula handling, visualization, and intelligent data import with semantic mapping.

Comprehensive toolkit for building AI data processing pipelines with CocoIndex, enabling ETL workflows, vector embeddings, and knowledge graph construction.

Specializes in managing context for multi-agent AI systems, ensuring seamless coordination and state preservation across sessions and complex workflows.

Specializes in configuring, implementing, and optimizing Model Context Protocol (MCP) servers for seamless integration within LLM-based AI workflows and applications.

Optimizes SQL queries, indexing, and caching for improved database performance in PostgreSQL and MySQL.

Evaluates AI bias, implements fairness measures, and ensures AI systems align with ethical standards and human values.

Diagnoses and resolves CockroachDB test failures, particularly Key-Value (KV) layer issues, to improve database reliability and debugging efficiency.

Enables integration of speech recognition, text-to-speech, and voice interaction features into AI applications with multi-provider support and real-time processing.

Builds AI chat agents for ArXiv paper exploration, enabling retrieval, content processing, and research synthesis via RAG.

AI-powered extraction of key decisions, action items, and patterns from meeting discussions.

Processes spreadsheet data including creation, editing, analysis, and transformation of Excel files with formula and formatting support.

Builds autonomous game-playing agents using reinforcement learning for strategy development, decision-making, and performance optimization in game environments.

Extracts and parses article content from web sources, retrieving text, metadata, and structured data while preserving formatting and context.

Guides users through LLM fine-tuning for domain-specific applications, covering data prep, training, and evaluation techniques.

Builds natural language interfaces for querying documents (PDFs, GitHub, emails) using RAG-based AI to extract and present information.

Coordinates specialized agents to solve complex problems through communication, task delegation, and result aggregation.

Builds RAG applications by integrating LLMs with external knowledge sources using vector databases and retrieval strategies for document Q&A and knowledge base systems.

Provides real-time discovery and tier classification (fast/default/heavy) of AI models from providers for accurate API integration.