Engineering

AI Engineering

Open-source tooling, agent skills, and competition projects at the intersection of AI, knowledge graphs, and intelligent systems.

Multi-Agent System · Databricks · Healthcare Data

Live 3D demo

Multi-Agent Crew for Automated Bronze-to-Silver Mapping

A 15-agent system that automates bronze-to-silver data mapping on Databricks, from table discovery to a reviewed, production-ready notebook with data-quality checks. Each of eight pipeline stages runs a small crew (worker, deterministic monitor, cross-model reviewer) with exactly one repair round, and agents coordinate only through three session stores, never by calling each other.

Multi-Agent SystemsLLM OrchestrationDatabricksPython

View Project

Agent Skills · Cursor · Databricks Genie

Enterprise Data-Engineering Skills for AI Agents

A governed library of enterprise-grade agent skills that turn recurring data-engineering work into reliable, automatable task playbooks. The same skills drive agents across Cursor and Databricks Genie — covering data ingestion, validation, quality checks, complex-file parsing, interoperability / data-sufficiency checks, EDI parsing, and transformations.

Agent SkillsTask AutomationCursorDatabricks Genie

View Project

RAG Architecture · Databricks · Organization Level

Organization-Level RAG over Data-Engineering Knowledge

A Databricks-native Retrieval-Augmented Generation system over internal data-engineering documents — field-mapping workbooks, data dictionaries, and design specs. It answers questions three ways (exact SQL lookups, hybrid vector retrieval, and LLM-drafted cited answers), fed by a governed ingestion pipeline with ~27 hard quality gates that block a bad build before it ever reaches answers.

RAGDatabricksVector SearchUnity Catalog

View Project

Remote MCP · Agent Infrastructure

Central AI Agent Hub with MCP for Enterprise-Scale Skill Deployment

A centrally managed MCP server that turns organizational AI skills into shared, versioned agent capabilities across teams and devices. It reduces duplicated prompt engineering, enforces consistent delivery standards, and gives a single deployment path for updates, tracking, and governance.

MCPModel Context ProtocolTypeScriptNode.js

View Project

World Finalist · Microsoft · Earth Category

Imagine Cup 2021

AI-driven environmental sustainability solution that reached the World Finals of the Microsoft Imagine Cup (Earth Category), combining computer vision with real-time sensor data.

Machine LearningComputer VisionMicrosoft AzureAzure IoT Hub

View Project

Open Source · AI Coding Assistant Skill

Reduce AI Token Use in Large Codebases

An open-source AI coding assistant skill that turns any folder of code, docs, papers, images, or videos into a queryable knowledge graph. Personally tested and validated — reduces token usage by ~50% in large codebases. Works in Claude Code, Cursor, Codex, Gemini CLI, and more.

Knowledge GraphAST / tree-sitterLeiden ClusteringNetworkX

View Project

More projects coming soon