Master's thesis · Machine learning
AI Strawberry Disease Detection
A hybrid YOLOv9 + DETR model that flags strawberry disease and ripeness with about 90% accuracy across 9,000+ images. I owned the full pipeline, from data preparation to evaluation.
Open to software & data roles
Amir Ghorbani · Software & Data
I build software and make sense of data.
Auckland-based developer with a Master of Computing, working across React frontends, IoT data pipelines and computer vision.
01Selected Work
Projects I'm proud of
Recent work across software, data pipelines and machine learning.
Data engineering · Databricks
NZ Electricity Lakehouse
128 million half-hourly wholesale prices since 1996 in a bronze → silver → gold Databricks lakehouse, with incremental loading, data-quality rules, an automated job and a dashboard. When the hydro lakes run low, power costs three times as much.
Data engineering · Live data
NZ Earthquake Analysis
175,000 GeoNet quakes modelled in SQL and Python, with a live API feed. Three in four are aftershocks, and a 2012 change in how magnitudes were measured would fake a calmer New Zealand.
Data analysis · Public data
NZ Rental Market Analysis
33 years of MBIE rental bond data modelled in SQL. Rents have stalled and, after inflation, sit 8% below their 2022 peak. I tested a supply hypothesis and rejected it.
More projects
02What I Do
What I do
01
Software Engineering
Responsive React and Next.js frontends backed by Node.js services and AWS, built from reusable components with performance and accessibility in from the start.
02
Data Pipelines & Analytics
Collecting, storing and moving data, from live sensor streams over MQTT and Node-RED to SQL databases, then turning it into dashboards people can read at a glance.
03
Machine Learning & Computer Vision
End-to-end ML in Python and PyTorch: preparing datasets, training and evaluating models. My thesis combined YOLOv9 and DETR to reach about 90% accuracy on 9,000+ images.
03About
A developer who works across software, data and machine learning.
I'm based in Auckland with a Master of Computing. I build clean, practical software, and I'm just as comfortable on the data side: designing pipelines, working in SQL and Python, and making results easy to understand.
My research background is in machine learning and computer vision, where my object-detection thesis took me through the full pipeline from data preparation to model evaluation. Whether the role is software, data or ML, I care about results that are accurate, fast and genuinely useful.
Based in
Auckland, New Zealand
Education
Master of Computing
Focus
Software, data & machine learning
Status
Open to software & data roles
04Tech Stack
Tools I work with
Software
JavaScript
TypeScript
React
Next.js
HTML
CSS
Git
Node.js
AWS
Data & Machine Learning
Python
SQL
PostgreSQL
MySQL
PyTorch
Machine Learning
Computer Vision
MQTT & Node-RED
AI Tools
Claude AI
ChatGPT
05Contact
Let's build something thoughtful, useful, and visually sharp.
I'm open to software and data roles, freelance projects, or anything interesting. Tell me what you're working on. I read every message and reply personally, usually the same day.