By skyforbes I’m currently running an experiment to see what modern AI (Gemini 3 Ultra / GPT-5 class) is actually capable of when acting as a de facto Senior Developer for someone who "can't write code."
The Context
I have an English degree and no CS background. I’ve been working in quality/production roles in the North East of England (currently a "Project Engineer" at a small manufacturing firm). My trajectory has been "replaceable drone" paid by the hour. I want to change that to "high-leverage problem solver" paid for outcomes.
The Problem
My current workplace runs entirely on Excel indexes and manual data entry. My specific role involves creating Manufacturing Route Cards—documents that tell the shop floor how to make a part (e.g., "Make 10x 2-inch nuts using this specific stainless steel bar"). This process involves manually scraping data from production drawings and typing it into Word templates. It’s slow, error-prone, and soul-crushing.
The Project (My "Portfolio Piece")
I realized that if I could code, I could automate this job out of existence. Since I can't code, I'm using AI to build a custom Odoo ERP module running on Docker to do it for me.
Here is the architecture I've built so far (with AI holding my hand):
- The ERP-Lite Backbone: A custom Odoo module that tracks the workflow from Sales Enquiry \rightarrow Proposal \rightarrow Work Order. This replaces the disparate Excel sheets and creates a "Single Source of Truth" for the admin team.
- The "Drawing Registry": A database of PDF production drawings. Crucially, this links the drawing to specific Engraving Logic.
- The Challenge: Metal parts need specific engravings (Heat No, Material Grade, Drawing Rev). This logic is complex and changes per client. I'm building a rules engine that pulls this data automatically from the material certs.
- The Route Card Generator: The "Money Tool." It takes the sales input, grabs the correct drawing from the registry, applies the engraving logic, and generates the PDF route card instantly.
- Dimensional Reports (The "Stretch Goal"): Originally I wanted to use OCR to read dimensions off the drawings for QC reports. Reality Check: OCR is too unreliable for engineering drawings (mistaking a Ø for a 0).
- The Pivot: I'm now building a system where "Critical Dimensions" are manually entered into the Registry once, then forever automapped to QC sheets. Humans do the thinking; the system does the remembering.
The Goal
I am trying to prove that a domain expert (me) + powerful AI can equal a "Systems Architect." I want to use this project to leverage myself into a role like Process Improvement Specialist or Odoo Functional Consultant, getting out of the "time-for-money" trap and into high-value systems work.
Discussion
- Has anyone else successfully pivoted from "Blue Collar/Admin" to "Tech/Systems" using a similar portfolio approach?
- Experienced Odoo/ERP devs: Am I over-engineering the Drawing Registry, or is this the right way to handle complex compliance data? Any advice/comments from the tech experts is welcome (you can be ruthless if you like)
- Thoughts on the "No-Code Architect" reality? Is this a viable career path, or will I hit a wall without deep syntax knowledge?
Any advice, critiques, or maybe let me know if you're working on something similar and don't consider yourself a cs expert. Quite happy for you to give me a reality check too, because of course ai is telling me this is absolutely possible and I'm the best person in the world for even thinking about this…. 🤔
(Yes my post was edited with an ai prompt for clarity and efficiency)
TLDR: Monkey wants big banana salary but lacks coding skills necessary. Monkey decides to use AI to help build automation processes. Is this a valid strategy to escape the UK wage trap, or does the monkey need a reality check?