Work experience

• Machine Learning and AI Engineer

  Independent Consultant

  2023.01-Current

  (3 years)

  • Built a geolocation application with a Qdrant vector-search backend using CLIP image features for location-aware retrieval. • Fine-tuned and deployed Qwen-based models for local training and serving, using Unsloth and LLVM-based optimization workflows. • Designed and implemented agentic AI workflows for task orchestration, tool use, and multi-step reasoning. • Designed industry management software for client-to-business workflows using Bolt.new.
• Machine Learning and Computer Vision Engineer

  Telexistence Inc.

  2024.07-2025.08

  (a year)

  • Integrated machine learning models into robotics perception and automation pipelines. • Optimized and refined model performance through data selection, model selection, and iterative evaluation. • Developed augmentation pipelines using synthetic and virtual data to improve robustness across operational scenarios. • Supported robotics ML deployment workflows by improving data quality, model reliability, and integration readiness.
• Machine Learning and Computer Vision Engineer

  Imagr Ltd

  2018.09-2022.09

  (4 years)

  • Led the design, training, and deployment of a real-time machine learning model for the Nvidia Jetson. • Developed a fast and cost-efficient classification architecture capable of accurately classifying over 10,000 categories, including challenging classes and out-of-distribution classification. • Conducted extensive experiments on classification models using Google Cloud TPUs to optimize hyperparameters, model architecture, and data selection, resulting in a highly performant and cost-efective solution. • Created an LSTM object detection model specifically designed to detect moving objects and implemented a tracker to track multiple objects over time, showcasing expertise in object detection, segmentation, and classification using raw Bayer data in low-light environments. • Maintained and deployed various machine learning models, including those for classification, clustering, object detection, segmentation, and barcode reading. Continuously monitored and optimized the performance of these models to ensure their ongoing success.
• Research Assistant

  Auckland Bioengineering Institute, New Zealand

  2017.03-2018.08

  (a year)

  • Built medical imaging tools for liver surgical planning, including CT segmentation, visualization, and shape modelling.

Educational experience

• University of Auckland

  BioEngineering

  2015.01-2016.09

  (2 years)

  * speed up finite element (FEM) with CUDA * augmented reality for surgery with oculus and stereo camera * pre surgical planning
• University of Auckland

  Engineering Science

  2012.02-2015.10

  (4 years)

  Mathematical modelling specialized for continuum mechanics

Languages