IoT · home lab · local-first AI · Tampere

I build things that scratch my own itch — local-first, data-sovereign, occasionally over-engineered.

IoT at the grid edge by day, home lab architect by night. Based in Tampere, Finland.

See What I'm Building

About Skadi Labs

I'm Charles Leerink. Skadi Labs is my personal R&D platform — the place where side projects live and occasionally become real things.

By day I'm Technical IoT Lead at Landis+Gyr, working on smart metering and the systems that keep the grid edge honest. Before that: eight years at Sonos, a stretch at Wirepas doing IoT mesh networking. The throughline has always been the same — bridging hardware, software, and the messy operational reality in between.

The projects here tend to rhyme. Local-first, because I'd rather own the database than rent access to mine. Simple and maintainable over clever and complex. Python and FastAPI as the default, because getting from idea to running service in an evening matters more than picking the most interesting tool.

Writing

Architecture decisions, implementation details, and things I figured out the hard way. Mostly about systems that run on hardware I own, AI that doesn't phone home, and the gap between how things are supposed to work and how they actually do.

Find more on my LinkedIn profile.

Playing in the ChatGPT Sandbox: A Technical Deep Dive

Playing in the ChatGPT Sandbox: A Technical Deep Dive

December 14, 2025 8 min read

What I discovered by treating ChatGPT like a Linux system and exploring the containerized environment behind every conversation

Read article →
What Happened After Giving SpinStack to Some Testers

What Happened After Giving SpinStack to Some Testers

October 25, 2025 7 min read

The bugs, fixes, and lessons from real-world testing of my self-hosted Sonos scrobbler.

Read article →
SpinStack Limited Beta: Your Music Data, Your Rules

SpinStack Limited Beta: Your Music Data, Your Rules

October 2, 2025 6 min read

After three months of dogfooding, SpinStack is ready for a small group of testers. A self-hosted music analytics platform for Sonos users who want to own their listening data.

Read article →
View all articles →

What I'm Building

Side projects that started as a problem and refused to stay small. Some are in production. Some are in the container graveyard. These are the ones still running.

In Production

SpinStack

Self-hosted music tracking platform. Monitors Sonos devices, logs every track played, and manages vinyl collection — without sending anything to Last.fm.

Python FastAPI SQLite
R&D Phase

OpenClaw

Local AI agent architecture with a tiered approach to RAG — routing queries through layers of specificity before hitting vector search, because standard RAG gets noisy at scale.

Python Ollama SQLite-vec
Active

The Shire

Middle-earth themed home lab. Hardened jump server and bastion host, configured with Ansible. SSH hardening, Fail2ban, AIDE file integrity monitoring. Next up: full observability stack.

Ansible Docker Nginx
Active

MökkiWatcher

Scrapes Finnish property listing sites and alerts on new mökki matches. Built because manually refreshing listings felt like playing the stock market.

Python FastAPI
Live

takeoutday.org

Countdown timer and instructions for downloading your data from the platforms that have it. Pure HTML, CSS, and JS — no frameworks, no trackers.

HTML CSS JS
Weekend Project

IKEA Price Checker

Compares IKEA product prices across Finland, Sweden, and the Netherlands with currency conversion. Rate limited, because IKEA's servers deserve respect too.

Python FastAPI

How I Can Help

I take on a small amount of consulting work where my background is actually useful — critical infrastructure, IoT integrations, and local-first AI systems. If that sounds like your problem, get in touch.

Technical Operations & Mission-Critical Systems

Systems where failure has real consequences

I've built technical operations for high-trust environments — energy infrastructure, IoT deployments, carrier integrations. Not strategy decks: actual implementation of hardened processes, incident response frameworks, and resilience systems that hold up when things go wrong.

What I bring:

  • Hardened process design for mission-critical operations
  • Technical integration work bridging hardware, software, and carriers
  • Security-of-supply architecture for critical infrastructure
  • ISO27001-informed incident response and operational resilience
  • Clear technical communication for complex stakeholder environments
Best fit for
  • Critical infrastructure operators (energy, utilities, industrial IoT)
  • Organizations integrating hardware, software, and carrier operations
  • Companies building high-trust technical operations from the ground up
Not a good fit for
  • Basic IT operations or standard SaaS deployments
  • Consumer-facing systems without regulatory requirements
  • Short-term firefighting (I build systems, not temporary fixes)

Local-First AI & Automation

AI that runs on your hardware, not someone else's

I build AI systems that stay on your infrastructure. Local LLM deployment, vector search pipelines, and practical automation — for environments where sending sensitive data to a cloud API isn't an option.

What I bring:

  • Local LLM deployment via Ollama for data-sensitive environments
  • RAG pipelines that run on your hardware with sqlite-vec
  • Automation workflows with proper access controls and audit trails
  • Technical documentation systems powered by AI without cloud dependencies
  • Working implementations, not just architecture recommendations
Best fit for
  • Organizations with strict data residency or sovereignty requirements
  • Critical infrastructure operators who can't use cloud AI
  • Teams wanting practical AI tooling without the compliance risk
Not a good fit for
  • Generic chatbot or customer service automation
  • Organizations fine with cloud-only AI solutions
  • Greenfield projects with no existing infrastructure to build on

Technical Training & Integration Enablement

Getting teams to actually use what you've built

I've delivered SDK coaching, API onboarding, and integration workshops across Europe, Asia, and North America — mostly for teams where hardware, software, and carrier operations meet and nobody fully understands the other side.

What I bring:

  • Integration training for hardware/software/carrier boundary problems
  • Technical workshop design for mission-critical deployments
  • SDK and API documentation written for the people who actually integrate
  • Cross-cultural delivery — 15+ countries, utility and industrial sectors
  • Field application engineering perspective from grid-edge deployments
Best fit for
  • Critical infrastructure vendors launching integration programs
  • Organizations bridging OT/IT for industrial or utility deployments
  • Teams that need technical enablement for high-trust stakeholders
Not a good fit for
  • General corporate training or soft skills workshops
  • Consumer-facing product education
  • Ongoing program management (I design and launch, not run indefinitely)

Get In Touch

Email is best. If it's consulting-related, a sentence or two about the problem is more useful than a brief.