Infrastructure
Compute, networking, storage, and private services across the Mac Mini M4, GB10, Pi nodes, and supporting systems.
Systems Portfolio
Applied AI systems running on a personal homelab
This work ties together local infrastructure, model-serving, automation runtimes, and operator-facing apps. The through-line is simple: build systems that are useful under real conditions, not just interesting in isolation.
local inference
AI orchestration
automation runtime
observability surfaces
operator-facing apps
Running across a Mac Mini M4 control plane, a Dell GB10 for local inference, and a cluster of Raspberry Pi nodes.
Mac Mini M4 control plane online
GB10 inference host active
Pi services powering automations
Models, dashboards, and apps in daily use
One system, multiple layers
The projects here are connected. Infrastructure supports the runtime, the runtime routes work across the stack, and the applications sit on top of that foundation as things I actually use day to day.
Runtime
Schedulers, model-serving, health checks, workflows, and the automation surface that keeps things moving.
AI layer
Local models, routing logic, agent orchestration, and tool-enabled execution across the environment.
Applications
Practical tools built on top of the system: dashboards, OCR workflows, decision support, and creative utilities.
Core systems powering the stack
These are the systems everything else depends on. If these layers stop working, the rest of the stack does too.
01
applications
Active System (v1 Live)
Application layer
This is the orchestration layer that keeps the environment operational — coordinating recurring jobs, health checks, alerts, and agent-triggered workflows across the stack.
Control plane
Mac Mini M4
Routing, orchestration, coordination, and system glue.
Inference
GB10
Local model serving, heavier workloads, and AI execution.
Services
Pi cluster
Always-on utilities, supporting apps, and private infrastructure.
Interfaces
Apps + dashboards
Status surfaces, workflows, tools, and the public-facing layer.
02
core infrastructure
Live Environment
Homelab Infrastructure
This is the foundation underneath everything else — local compute, networking, storage, and the private environment that makes experimentation reliable instead of fragile.
03
ai systems
Active Development
AI Agent Orchestration
This is the AI layer that makes local models operational — routing work, invoking tools, and handling real multi-step workflows across machines.
04
applications
Active System (v1 Live)
Application layer
This is where the system becomes tangible — applications that consume shared infrastructure, automation, and local AI capabilities to solve real problems in day-to-day workflows.
Selected applications
A curated set of operator-facing applications built on top of shared infrastructure, automation, and local AI — spanning sales intelligence, monitoring, OCR workflows, creative tooling, and personal automation.
Active System (v1 Live)
Co-Worker
AI-powered sales intelligence system that turns daily signals into account insights, prioritization, and actionable outreach workflows.
Active System (v1 Live)Cronitor
Runtime monitoring system for tracking job execution, failures, and automation health across distributed workflows.
Active System (v1 Live)PiSplit
OCR-driven expense workflow that converts receipts into itemized splits, balances, and reimbursement tracking.
Active System (v1 Live)Culler
Local-first image triage workflow for rapidly reviewing, selecting, and preparing photos for editing.
Active System (v1 Live)Apartment Scout
Automated listing pipeline that collects, filters, and ranks apartments based on real-world search criteria.
Supporting systems and adjacent infrastructure
A broader view of the infrastructure, model-serving, and supporting systems that make the application layer possible.
Core Infrastructure
Compute, orchestration, model delivery, and reliability foundations that power every downstream system.
AI Systems
Agent runtimes, model routing, and intelligence services that turn models into practical operators.
This work reflects how I approach systems engineering: designing real infrastructure, running local AI, and building tools that are actually used.
Let’s talk