Jordan Lowe
self-taught builder · automation & applied AI
Self-taught builder who directs AI to turn everyday operational friction into working tools — the judgment to steer and verify each build is the real skill, not the syntax. Quality Control Inspector for commercial HVAC rooftop units at The Master Group: equally at home on the distribution-centre floor and in the codebase, and with technical and non-technical teams alike.
Experience
Quality Control Inspector — commercial HVAC rooftop units · The Master Group
2024 — present
- Designed, built, and shipped a Power Apps + Power Automate + SharePoint inspection platform running product & outbound QC in 3 distribution centres — self-initiated.
- Cut average inspection entry from 1:40 to 22 seconds (~80%) by pre-filling the passing path so inspectors triage exceptions — recovering 40+ hours/year across ~2,000 inspections.
- Wrote the documentation and trained teams at each centre — earning adoption from three groups with three different ways of working.
- Now developing the next iteration: an RF-DETR damage-detection model trained on the platform’s curated dataset, aimed at near-zero inspection overhead and consistent results across centres.
Data Analyst — Financial Analysis · Rim and Tire Co
2023 — 2024
- Prototyped an automated pipeline pulling public financial data into macro-trend models — Python (pandas, NumPy), Power BI, Tableau, Excel — with GPT-3.5 in the workflow back in 2023.
Earlier — operations, trades & field work
2015 — 2023
Fibre-Optic Installer · Granite Electrical — commercial fibre installation & service for business clients, customer-facing field work2022–23
Mechanical Sales · Lowe Mechanical Sales — translating technical product specs for customers2019–20
Plumbing Labourer · Centurion Mechanical & Cold Frog2018–19
Warehouse Management · Centurion Mechanical — receiving, inventory & staging2016–17
Produce Stocker · Walmart2015–16
Projects — all built by directing AI agents
QC Inspection Platform
The Master Group
The Power Platform app above — pre-fill-the-pass design from knowing 90%+ of units pass; automated storage, categorization & cleaning of every inspection into a training-ready dataset.
RTU Damage Detection
The Master Group
Detection transformer (RF-DETR, PyTorch) trained on our inspection data; active-learning labelling queue ranked by model disagreement; self-hosted, serving live inference on field photos.
Mînî — database engine from scratch
Personal
~52,000-line C++ database engine — transactions, crash recovery, a query optimizer — every piece verified: 110/110 test binaries green under sanitizers.
Îktomnî — knowledge engine & research agent
Personal
Go document-intelligence engine: PDF → knowledge graph (Neo4j) → cited answers, plus an agentic deep-research loop — running entirely on local models (Llama 3.1 8B) on my own hardware.
Skills
DIRECTING AIOrchestrating coding agents, prompt & context design, verification & eval, rapid iteration
MICROSOFTPower Apps, Power Automate, SharePoint, Power BI, Excel, Word, Outlook
DATA & CODEPython (pandas, NumPy), SQL, Tableau, data analysis · Go, C++ (AI-assisted) · ML & computer vision (PyTorch)
FOUNDATIONSProcess mapping, documentation & training · Windows / Linux / macOS, command line, web development