Open-source project · GitHub01
Promptboard
Promptboard is a local-first agentic development environment for building software with AI coding agents.
Agentic AIOpen SourceLocal-first
What I’m working on
A space for my personal projects, research and articles.
Tutorial02
Remote Control Layer for AI Coding Agents
How I use Herdr, Tailscale and Termius to organize parallel AI development, keep long-running agent sessions alive, and supervise the whole environment remotely from an iPhone.
Agentic AIHerdrRemote Development
Article03
Secure ML Collaboration: Sharing a Local Nvidia GPU Workstation with Zero Trust
How I used Twingate and Zero Trust Network Access to securely share a Linux GPU machine for a collaborative machine-learning project.
Zero TrustSecure MLGPU Infrastructure
Article04
LLM Burners: Hardwiring AI Models Into Silicon
Modern AI inference has a surprisingly physical bottleneck: moving model weights around is expensive.
AI HardwareLLM InferenceTaalas & AMD
Open-source project · GitHub05
QuantJourney Backtester
Open-source, evidence-first backtesting for systematic trading research—turning strategy ideas into auditable, reproducible research.
PythonBacktestingML & Data Engineering
Tutorial06
Use Subscription-Authenticated Cloud LLMs in DeepSeek Harness
Connect an authenticated Codex runtime to DeepSeek Harness through a local OpenAI-compatible proxy—without configuring a separate OpenAI Platform API key.
DeepSeek HarnessCodexCloud LLMs
Article07
The A2A Protocol and the Agent-to-Agent Economy
Following my AI Mesh Networks article, this explores how autonomous agents discover capabilities, establish trust, delegate work, transact and form temporary supply chains.
A2A ProtocolAgent EconomyAgentic Infrastructure
Article08
From Model Intelligence to System Intelligence
How AI Mesh Networks, model and agent routing, A2A and MCP are moving AI architecture from isolated models toward coordinated system intelligence.
AI MeshModel RoutingA2A
Article09
What Is an AI Agent Harness?
A systems-engineering look at the execution and control layer around AI models—and why reliable context, tools, state, verification and security determine whether agents actually work.
AI AgentsHarness EngineeringAgent Architecture