Skip to content

Harry Lo

Hi, I’m Harry. I’m a software engineer based in New Zealand.

This website is my personal space for documenting what I learn, build, experiment with, test, research, and gradually understand.

I explore AI engineering, agentic software development, cloud technologies, data, financial market technology, investing, algorithmic trading, home lab systems, networking, and practical software projects.

This is not a finished encyclopedia. It is an evolving engineering notebook, project archive, and learning journal built from real hands-on work.

Practical notes, experiments, architecture decisions, project write-ups, and lessons learned all live here.

  1. Who I Am
  2. Why I Created This Website
  3. What You’ll Find Here
  4. Start Here
  5. Featured Articles
  6. How This Blog Works
  7. What I’m Building
  8. How I Use AI
  9. Projects as Learning Environments
  10. Where Should You Start?
  11. Follow My Journey
  12. About This Site

I am a software engineer with an engineering background and a strong interest in:

  • Software development
  • AI and AI-assisted development
  • Cloud technologies and infrastructure
  • Data and analytics
  • Home labs and networking
  • Financial markets and investing
  • Algorithmic trading
  • Technical experiments and personal software projects

A lot of what I learn comes from building real projects, solving practical problems, experimenting with new technologies, and documenting both what worked and what did not.

I prefer learning by building rather than only reading about a subject.


I created HarryLo.com as a place to organise and share my learning journey.

Over time, my notes and ideas started becoming scattered across:

  • Repositories
  • Chats
  • Documents
  • Bookmarks
  • Different apps and tools
  • Project folders
  • Research notes

I wanted one place where I could gradually turn those fragments into something more structured, searchable, and useful.

The goal is not to present myself as an expert in every subject.

The goal is to document the process of learning.

I want to record:

  • What I was trying to understand
  • What problem I was trying to solve
  • What I researched
  • What I built
  • Why I made certain engineering decisions
  • What worked
  • What did not work
  • What I would change next time
  • How my thinking develops over time

The site will gradually grow around several areas I am actively learning and building in, including:

  • Software development
  • AI engineering
  • AI-assisted and agentic software development
  • Cloud and infrastructure
  • Home lab and networking
  • Data and analytics
  • Markets and algorithmic trading
  • Trading-system architecture
  • Personal software projects
  • Computer vision and smart security
  • Astro and web development
  • Technical experiments and lessons learned

The articles on this site are not all the same type.

Some are:

  • Tutorials
  • Project notes
  • Architecture decisions
  • System-design notes
  • Research journals
  • Technical experiments
  • Explanations of concepts I wanted to understand more deeply
  • Records of a problem, the approach I tried, and what I learned from it

Together, they form a record of how my understanding develops.


If you’re new to the site, these are the main areas I’m currently exploring.

Learning notes, AI engineering roadmaps, model integration, RAG, agents, orchestration, and practical experiments.

This area includes both learning plans and hands-on experiments with building AI-enabled systems.

Experiments with Codex, Claude Code, Cursor, Grok, multi-agent workflows, automated testing, review, software-development orchestration, and AI-assisted task management.

The workflow is gradually evolving from:

Requirement
↓
AI Coding Agents
↓
Pull Request

toward a broader development system:

Idea
↓
AI Idea-to-Ticket Workflow
↓
GitHub Issue
↓
Multi-Agent Coding Workflow
↓
Pull Request
↓
Human Review

Related articles include:

Implementation repo: I am turning this workflow into code in the public ai-dev-workflow repository.

Done: Stage 1 — repo foundation, agent entry points (AGENTS.md / CLAUDE.md), architecture and setup docs, and a clear public vs private boundary. This site documents the design in the Agentic Software Development article series.

Next: Stage 2 — idea-to-ticket MVP (local ticket schema, validation, fixture CLI, then one live AI path for a single ticket). After that: GitHub integration (Stage 3), multi-agent coding MVP (Stage 4), then connecting both pipelines (Stage 5).

Market data research, options concepts, trading-system architecture, stock screening, momentum modelling, algorithm research, and technical analysis from a software-engineering perspective.

This is where I connect software engineering with financial-market research.

Home servers, networking, firewalls, storage, self-hosting, infrastructure, security, and hands-on system experiments.

Experiments with ANPR, Frigate, computer vision, cameras, detection pipelines, and practical smart-security systems.

Notes from building this website with Astro and Starlight, including multilingual content, documentation structure, SEO, information architecture, and practical web-development patterns.



Most of the content on this site follows a simple pattern:

Problem
↓
Research
↓
Experiment
↓
Build
↓
Learn
↓
Document

When I encounter something I do not understand, I try to turn it into a small project, experiment, or article.

Sometimes the result is a working feature.

Sometimes it is a failed experiment.

Sometimes the main result is simply a better understanding of the problem.

All three outcomes can still be useful.

Over time, these individual pieces become a record of how my technical thinking develops.


A lot of the content on this site grows out of real projects rather than isolated tutorials.

I am experimenting with:

  • Multi-agent AI development workflows
  • Automated coding pipelines
  • AI-assisted ticket creation
  • Architecture and planning agents
  • Implementation agents
  • Automated testing
  • Independent AI review
  • Human approval gates
  • Software-development orchestration

The aim is not simply to ask AI to write code.

I am interested in how AI can participate in a controlled engineering process from idea to implementation and review.

Implementation progress lives in the public ai-dev-workflow repository (Stage 1 complete; Stage 2 idea-to-ticket MVP in progress).

My market-technology work includes:

This includes:

  • Remote roadside ANPR plan
  • Frigate proof-of-concept work
  • Camera and detection architecture
  • ANPR research
  • Computer-vision experiments

My home-lab work includes:

HarryLo.com is also one of my learning projects.

I use Astro and Starlight to explore:

  • Multilingual documentation
  • Content organisation
  • Information architecture
  • SEO
  • Technical writing
  • Reusable components
  • Content automation
  • Documentation workflows

AI is becoming part of both my software-development workflow and my learning process.

I use it to help with tasks such as:

  • Exploring architecture options
  • Understanding unfamiliar concepts
  • Turning rough ideas into structured engineering tasks
  • Planning implementation
  • Reviewing code
  • Testing assumptions
  • Comparing different approaches
  • Documenting experiments
  • Organising research
  • Drafting and refining technical notes

But I do not want the workflow to become:

Idea
↓
AI immediately writes code

I am more interested in a controlled process:

Idea
↓
Research
↓
Decision
↓
Ticket
↓
Implementation
↓
Testing
↓
Review
↓
Human Approval

That development process itself has become one of the subjects I document on this site.


Some of the projects documented on this site include:

  • Software applications
  • AI-assisted development workflows
  • Data and trading research tools
  • Market-analysis experiments
  • Home-lab infrastructure
  • Computer-vision systems
  • Website and documentation tooling

The projects are not necessarily finished products.

They are working environments where I can:

  • Test ideas
  • Learn technologies
  • Compare approaches
  • Make mistakes
  • Measure results
  • Improve my understanding
  • Document the lessons

A project can be useful even before it becomes a polished product.

Sometimes the main value is the engineering knowledge created along the way.


If this is your first visit, start with whichever area interests you most.

You can:

  • Explore the site by topic
  • Follow one of the project series
  • Read the featured articles
  • Browse the latest technical notes
  • Follow a project from early experiments to later versions

There is no required reading order.

The site is designed to grow with my learning journey.

If you’re interested in AI-assisted development, the Multi-Agent AI Development Workflow is a good starting point.

If you’re interested in market technology, start with Market Data Research or DawnLuma.

If you’re interested in infrastructure, start with Home Server & Home Lab Design.

If you’re interested in how this site is built, start with Astro from Zero.


HarryLo.com is a long-term learning project.

As I:

  • Build new things
  • Learn new technologies
  • Change my mind
  • Discover better approaches
  • Find mistakes
  • Revisit old decisions
  • Improve existing projects

I will continue updating and expanding the site.

Some pages may change as my understanding improves.

That is intentional.

The site is meant to reflect an active learning process rather than a fixed snapshot.

Thanks for visiting.


This site is built with Astro and Starlight.

I also document parts of the site’s architecture and implementation in the Astro section.

HarryLo.com is both a publishing platform and a project in its own right: a place where I can practise engineering, documentation, organisation, and long-term knowledge building.