3. Data, Science & AI

Enables writing, reviewing, and debugging of Stata code for data cleaning, analysis, and documentation, supporting .do and .dta file operations.

Retrieves health metrics from Oura Ring wearable device via CLI, including sleep, activity, heart rate, and readiness scores.

Generates realistic synthetic datasets with non-linear distributions and integrity constraints using Faker and Spark, saving to Databricks for testing and demonstration.

Enables creation and management of Databricks Agent Bricks for conversational AI applications, including document Q&A, SQL exploration, and multi-agent orchestration.

Creates and manages Databricks Lakeflow Spark Declarative Pipelines (SDP/LDP) for data processing, including streaming, CDC, and medallion architecture patterns.

Provides evaluation utilities for GenAI agents using MLflow 3, including trace analysis, dataset building, scorer creation, and execution.

Creates synthetic PDF documents for RAG systems and unstructured data testing, enabling evaluation datasets and demo content.

Enables discovery, implementation, and management of EachLabs AI provider integrations and API status updates.

Enables discovery, implementation, and documentation updates for Runware AI models and their API providers.

Connects Streamlit applications to Snowflake for database access, secret management, and SQL data querying in data-driven applications.

Builds data-driven dashboards with Streamlit for KPI displays, metric cards, and responsive layouts in data science contexts.

Visualizes data through charts, dataframes, and metrics in Streamlit, supporting native charts, Altair, and column configuration.

Command-line interface for interacting with large language models, enabling transcription, vision analysis, speech processing, and image generation.

Best practices for Bluera Knowledge store lifecycle: creation, indexing, management, and failure handling.

Enables Bayesian statistical modeling with PyMC v5+, including MCMC inference, model diagnostics, and advanced techniques like hierarchical models and Gaussian processes.

Synthesizes insights from multiple AI models (Claude, Gemini, Codex) to analyze complex problems, identifying consensus, disagreements, and blind spots.

Optimizes AI agent routing through LLM-guided context building and traceable decision logic for enhanced system accuracy.

Designs and modifies RAG pipelines using LangGraph, hybrid retrieval, GraphRAG, and LLM prompt engineering.

Expert in managing .faf (Foundational AI-context Format) files for persistent AI project context, including AI-readiness scoring and MCP server configuration.

Adds a direct name mapping between Steam game titles and How Long to Beat (HLTB) database entries for accurate data alignment.

Automates data extraction from App Store Connect, mini-programs, and Xiaohongshu for COO operational analytics.

Assists in developing autonomous agents for Claude Code plugins, covering structure, system prompts, triggering conditions, and best practices.

Enables vector-based semantic search retrieval for Agent-to-Agent communication in AI systems.

Manages vector collections for A2A agents, enabling efficient storage and retrieval of vector embeddings in agent communication systems.