Book an audit
About Liam Penkethman

Self-taught, output first, and accountable to my own customers.

I did not come through a computer science degree. I came through eight years of Shopify development, fifteen years of aquascaping, and a habit of finishing difficult builds. The credibility is in the output.

The through line

Deep domain knowledge in a physical craft, paired with production software.

Few people combine biological domain knowledge with production-grade engineering. That combination is the reason my e-commerce work lands differently. I understand stock, fulfilment, refunds, seasonality and the customer waiting on a late parcel, because I carry all of it in my own store.

I build tools I use myself first, then package them for others. Nothing on the services list is theoretical.

I am also finishing a Certificate II in Aquaculture at Fremantle Maritime TAFE, formal training layered on fifteen years of hands-on work.

Operating rules
  1. Production-ready by default

    Modular, scalable architecture. No prototype handed over as a product.

  2. Confidence gating on generated data

    Auto-generated output below 0.90 confidence gets held, not published.

  3. Stability over a quick sale

    I do not recommend an incompatible combination because a sale sits on it.

  4. You own everything

    Code, repository, accounts and documentation transfer at handover.

The stack daily use

What I work in.

Tools are not what you buy. Listing them saves you a discovery call asking whether I have touched your platform.

Languages and runtime

TypeScriptJavaScriptPythonNode.jsReactNext.jsLiquid

Data and infrastructure

PostgreSQLPrismaRedisDockerCI/CDRailwayVercelAWSCloudflare

Commerce and integration

Shopify Admin APIStorefront APIMetafieldsTheme extensionsStripeRESTGraphQL

AI systems

RAG architectureVector databasesTool callingAgent orchestrationClaude APIElevenLabs

Pipelines

Web scrapingETLWorkflow automationBullMQSocket.io

Domain

Live-goods logisticsWater chemistryCompatibility modellingRetail operations
How I sell

I do not sell technology because it sounds impressive.

I start with the problem. What costs you time. Where customers drop out. Which work keeps repeating. Which information sits somewhere nobody reaches.

Then I pick the simplest practical combination of design, software, automation and AI that produces a commercial result. Sometimes the answer is a rebuilt product page, not an AI system.

I use modern development tooling to move faster than a solo operator normally does. The tooling is not the product. The outcome is.

Next step

Start with a call. No pitch deck.

Tell me what is costing you time or sales. If an audit is the wrong next step I will say so.