Devon Copp

Machine Learning Engineer in Auckland, New Zealand. Devon owns machine learning and AI projects end to end, from framing the problem to shipping the system around the model.

• Email: devoncopp@gmail.com
• Website: https://devoncopp.com
• GitHub: https://github.com/djcopp
• LinkedIn: https://www.linkedin.com/in/devon-copp-35a449177/
• Full CV: https://devoncopp.com/cv
• Full CV as Markdown: https://devoncopp.com/cv.md

What Devon works on

• Models: designed and trained for the problem, from vision to graph neural networks.
• Systems: deployment, MLOps and agent tooling, on device or in the cloud.
• Product: built for the people who use it, with care for the design.

Work

ZOZO New Zealand, Machine Learning Engineer (2023 to now)

The Auckland team behind ZOZOFIT, building measurement technology for ZOZO in Japan. Devon owns models from data and training through to the phone.

• Body scanning on the phone: trained and deployed the deep learning vision models behind the ZOZOFIT scan, which measures a body from photos taken on a phone. The result is a faster, steadier scan in production.
• Other areas: computer vision, recommendations, MLOps, and agentic AI.

Halter, Graduate Core Systems Analyst, then Data and Systems Analyst (2021 to 2023)

Smart collars that let farmers fence, move and watch their herd from an app. Devon joined as employee number 64. Halter is now a $2 billion company.

• Algorithms for cow safety: built the detection algorithm behind a patented herd-wide lockout. It spots an obstruction from the first few cows to reach it, and the collars tell the rest of the herd over Bluetooth. Built with a small team of engineers.
• Power for a 24/7 wearable: developed behaviour detection and power algorithms in the firmware of a solar-powered collar, using lightweight signal processing to switch components off when they aren't needed, with full function and safety kept.

Projects

• autobaseline: a Python CLI that takes a dataset and a target metric, then launches Claude Code to scaffold a full ML project and run an autonomous keep-or-discard experiment loop. https://github.com/djcopp/autobaseline

Education

Bachelor of Engineering in Engineering Science (First Class Honours), University of Auckland, 2017 to 2020.

How Devon works

1. Frame: find the real problem with the people who have it.
2. Measure: build the yardstick before the model.
3. Iterate: change one thing, measure, keep what wins.
4. Ship: deploy, monitor and hand over a system that stays good.

Contact

Email is the best way to reach Devon about a project: devoncopp@gmail.com. No phone number is published.

Connect over MCP

This profile and the full CV are also available from a read-only MCP server at https://devoncopp.com/mcp, with no key needed. Tools: get_profile and get_cv.