◐ Shell
clean mode source ↗

orneryd - Overview

Hi, I'm TJ Sweet

Principal / Staff Software Engineer focused on AI infrastructure, databases, open source systems, and enterprise platforms.

I build infrastructure other engineers build on: graph/vector databases, high-performance UI platforms, Web Components, telemetry tooling, developer frameworks, and production systems that need to be fast, reliable, and maintainable.

GitHub LinkedIn

What I work on

  • AI infrastructure for agentic systems, Graph-RAG, long-horizon memory, embeddings, local LLM inference, and retrieval platforms
  • Graph, vector, and temporal database systems with API-compatible surfaces and production-grade execution paths
  • Enterprise UI architecture with Angular, React, Web Components, Shadow DOM, and cross-framework design systems
  • Cloud-native platforms using Go, Rust, TypeScript, Python, Kubernetes, GCP, Azure, AWS, and observability tooling
  • Open source platforms, developer tools, framework migrations, and systems other engineers rely on

Engineering principles

  • Build boring, reliable foundations that scale beyond the first use case
  • Prefer clear APIs, compatibility layers, and migration paths over rewrites for their own sake
  • Treat performance, accessibility, maintainability, and developer experience as product features
  • Design systems that teams can understand, extend, operate, and trust in production
  • Contribute to open source when the solution should outlive a single company or team

Featured open source

NornicDB

Independent graph + vector + temporal MVCC database for AI agents, Graph-RAG, long-horizon memory, and canonical truth systems.

Highlights:

  • Neo4j-compatible Bolt/Cypher and Qdrant-compatible gRPC API surfaces
  • Unified execution path for graph traversal, HNSW vector retrieval, MVCC reads, embeddings, local LLM reranking/inference, GPU acceleration, MCP tooling, schema constraints, and policy-driven retention
  • Agentic memory architecture with decay profiles, promotion policies, canonical graph-ledger supersession, GDPR retention, and MVCC-aware scoring semantics
  • Local Apple M3 Max benchmark: vector_only p50 578us at 15,767 req/s; vector_one_hop p50 1.6ms at 6,081 req/s with 800 concurrent requests and 0 errors

uiGrid

Original author of AngularJS UI Grid and maintainer of the modern cross-platform remaster.

Highlights:

  • Original AngularJS UI Grid reached 5.4K+ stars and 2.4K+ forks
  • Modern remaster keeps the familiar gridOptions / columnDefs / onRegisterApi API surface
  • Rebuilt for Angular 21+, React 18+, Web Components, Rust/egui, and native C/LVGL
  • Includes virtualization, sorting, filtering, grouping, tree data, editing, keyboard navigation, selection, pinning, state persistence, and CSV/Excel/PDF export

Professional focus

Currently, I am a Staff Software Engineer in AI Labs at CVS Health, working on AI-powered tooling, platform modernization, standards-based application architecture, and enterprise-scale Angular/Web Components migration.

Previously, I have worked as a Principal Software Engineer, Lead Software Engineer, Principal Software Architect, founder, consultant, and open source platform lead across financial services, healthcare, SaaS, automotive, fraud systems, cloud platforms, and UAV hardware/software systems.

Some systems I have built or led include:

  • High-security SaaS services for asset-management and global-authentication workflows
  • Global Web Component UI frameworks adopted across internal enterprise services
  • DITA-based multi-tenant authoring and publishing platforms
  • Telemetry and CI/CD integrations for Kubernetes and cloud delivery pipelines
  • Experience APIs and insurance workflow UIs
  • Consumer UAV flight-controller products sold in 12+ countries
  • Build-Your-Own vehicle configurator platforms across major automotive brands