Lab · Projects
Built and operated.
Platforms and systems, with what they changed rather than what they contain.
- 2026MemoryOSLocal-first memory pipeline for macOS — captures screen, audio, email, calendar, and documents as structured Markdown for AI-powered querying. Reference implementation for Context Compilation Theory.
- 2026Praxo-PICOSLocal-first personal intelligence operating system for macOS. Captures context from email, calendar, screen, documents, and notes — makes it searchable and actionable. Privacy-preserving, self-healing, commercial-grade.
- 2026Praxo-AgentOSGoverned lifecycle tooling for agentic engineering work. Reference implementation for Cybernetic Software Delivery — nine-stage lifecycle, artifact taxonomy, four-tier governance, and DORA-extended metrics.
- 2026SemanticStudioAn open-source, data-first multi-agent chat platform with 28 configurable domain agents, 4 fully-customizable chat modes, 4-tier memory with Context Graph, GraphRAG-lite, self-learning ETL, and comprehensive observability.
- 2025Reinforcement Learning ETL PipelinesBuilding next-generation ETL pipelines powered by reinforcement learning with advanced Plan-Act-Reason loops that enable true self-learning capabilities—pipelines that reason about their transformations and continuously optimize their behavior.
- 20253D Ontological Data ArchitecturesConstructing three-dimensional ontological representations that capture the full complexity of enterprise data architectures—enabling AI systems like RAG and recommendation engines to navigate semantic relationships with unprecedented depth.
- 2025Attention-Optimized Context EngineeringAdvanced context engineering techniques that exploit transformer attention patterns to maximize information utilization in agentic systems—positioning critical context where attention naturally focuses.
- 2025Agentic RAG ArchitecturesDesigning and implementing production-grade agentic RAG systems with 12-15+ specialized agents that optimize retrieval pipelines across terabytes of enterprise data, overcoming recency bias and attention dilution through advanced context engineering.
- 2025Advanced Context EngineeringResearch and implementation of novel context engineering techniques that overcome fundamental limitations in attention mechanisms, enabling more effective use of extended context windows in production LLM applications.
- 2025Self-Learning Data PipelinesBuilding intelligent data integration systems that use Plan-Act-Learn loops to continuously adapt, learn, and optimize—moving toward fully autonomous data pipeline management.
- 2024-PresentEnterprise Data FabricDesigning and implementing enterprise data fabric architectures that unify disparate data estates, enabling trustworthy AI through comprehensive governance, lineage, and quality frameworks.
- 2025Real-Time Quality Feedback SystemsDeveloping production systems for real-time AI output quality assessment, including hallucination detection, factual grounding verification, and reinforcement learning-based response optimization.
- 2025Future of AI Operating SystemsResearch and published writing on the evolution of software interfaces toward AI-native paradigms, where traditional SaaS becomes headless infrastructure and agentic systems become the primary user interface.
- 2010-PresentAcademic Teaching & ResearchOver two decades of graduate-level teaching, doctoral mentorship, and academic research in AI, data architecture, and enterprise systems. Published author of books, academic papers, and industry courses.