Devon Copp

Machine Learning Engineer, based in Auckland, New Zealand.

• Email: devoncopp@gmail.com
• Website: https://devoncopp.com
• GitHub: https://github.com/djcopp
• LinkedIn: https://www.linkedin.com/in/devon-copp-35a449177/

Summary

• Machine Learning Engineer and Data Scientist with strong experience in the full end-to-end ML solution lifecycle: exploration, experimentation, modelling, deployment, and iteration.
• Strong analytical problem-solving: framing ambiguous problems with data-driven decision-making, and following through to get measurable customer and business impact.
• Experienced in algorithm design and evaluation, including statistical modelling, traditional machine learning, optimisation, deep learning, and agentic AI.
• Skilled in relationship management, with strong communication experience across customers, non-technical colleagues, and in supporting and mentoring junior engineers and interns.
• Maintains up-to-date knowledge of advancements in machine learning and AI, applying best practices across model development and software engineering.

Tech stack

• Software engineering: Python, SQL, AI-driven development
• Frameworks and tools: PyTorch, Scikit-learn, Docker, OpenCV, AWS / GCP, Claude Code/Codex, Tableau
• ML: Deep Learning Training, Traditional ML, Experimentation & A/B testing, Deployment, AI agents

Experience

Machine Learning Engineer, ZOZO NZ

August 2023 – Present

• Trained and deployed the deep learning vision models behind ZOZOFIT's body scan on the phone, delivering a faster, steadier scan in production.
• Built internal agentic AI tooling, including independent coding agents, memory systems, MCP servers, and agentic code review systems.
• Applied analytical problem-solving to a range of computer vision and recommendation problems, selecting appropriate modelling approaches based on data characteristics and product requirements.
• Led technical evaluation and collaborated on the migration away from expensive licensed software, analysing open-source alternatives and building conversion pipelines for historical data.
  • Developed a computer vision model to predict and project 3D body models, replacing parts of a traditional optimisation pipeline and improving speed by ~20%.
• Designed and trained lightweight deep learning models for on-device computer vision, reducing model size from 6M to 800K parameters through custom architecture design while maintaining business-critical accuracy targets, enabling on-device deployment.
• Developed and maintained internal ML training infrastructure supporting reproducible experimentation and model lifecycle management.
• Deployed multiple ML systems as Dockerised APIs, supporting concurrent users and enabling rapid testing by product teams.
• Led development and delivery of multiple machine learning systems, including:
  • Graph Neural Network recommendation systems learning user-to-content embeddings to improve product recommendations.
  • Deployed and demoed open-source models with product or system impact to business development teams and executives.
• Mentored junior engineers and interns through technical guidance, code review, and structured problem-solving. Represented the company at career events and recruiting initiatives.

Data and Systems Analyst, Halter

September 2021 – August 2023

• Built the detection algorithm behind a patented herd-wide lockout, spotting an obstruction from the first few cows to reach it so collars could alert the herd over Bluetooth, with a small team of engineers.
• Developed behaviour detection and power algorithms in the solar-powered collar's firmware, using lightweight signal processing to switch components off when idle while keeping full function and safety.
• Designed and evaluated algorithms supporting core product systems including navigation, communications, and power management of the Halter Collar.
• Developed Python prototypes for firmware and cloud-based algorithms, modelling their impact on product KPIs using historical operational data.
• Applied statistical analysis, optimisation, and machine learning techniques to solve engineering and logistics problems across the product ecosystem.
• Collaborated with product managers and engineering teams to prioritise projects using exploratory data analysis and technical insight.
• Led hiring and mentoring for data analysts and interns, including CV screening, technical interviews, and ongoing mentorship.

Graduate Core Systems Analyst, Halter

April 2021 – September 2021

• Joined as employee number 64; Halter is now a $2 billion company.
• Performed exploratory analysis on large operational datasets to identify product improvement opportunities.
• Designed and analysed A/B tests to validate product features in production environments.
• Built Tableau dashboards used by engineering, ML, and product teams to monitor performance and guide decision-making.

Projects

autobaseline

2026 – https://github.com/djcopp/autobaseline

• Built a Python CLI that inspects a dataset and target metric, then launches Claude Code to scaffold a full ML project: data loaders, baseline model, training, evaluation, FastAPI serving, and a Dockerfile.
• Runs an autonomous experiment loop after a data leakage audit: each change is committed, trained, and evaluated, and kept only if the target metric improves.

Education

Bachelor of Engineering in Engineering Science (First Class Honours), University of Auckland

2017 – 2020

Specialised in statistics, machine learning, and optimisation, with coursework in advanced machine learning, data science, and mathematical modelling.

Relevant coursework:

• Advanced Machine Learning
• Advanced Data Science Practice
• Integer & Multi-Objective Optimisation

Final year dissertation: Using Virtual Reality to Create Inter-Brain Synchronisation

• Developed a VR experimental environment in Unity (C#) for social interaction experiments using EEG hardware.
• Built a data pipeline to automatically process and analyse EEG signals collected during experiments.

Additional experience