3DMorphic

3DMorphic is an Australian medical device company that pioneered patient-specific spinal implants. These are custom titanium implants designed directly from a patient's CT scan and manufactured via 3D printing. Over my five years with the company, I worked across software engineering, production engineering, and applied machine learning to scale their operations.

Role Context

Standard spinal implants are mass-produced in fixed sizes, but complex anatomies or severe deformities require a perfect, custom fit to reduce revision rates. Producing these devices required a seamless software-to-hardware pipeline: segmenting the CT scan, designing the implant in CAD, validating it mechanically, and manufacturing it in titanium. My role was to eliminate bottlenecks across this entire lifecycle by building specialized software tools and robust IT infrastructure.

Core Engineering Achievements

Custom CAD Application and Geometry Engine

The legacy workflow required a design engineer to manually sculpt implant geometry in general-purpose CAD tools, taking over a week per implant. I architected and built a purpose-built CAD application from scratch using modern C++. Additionally, I engineered a high-performance C++ geometry engine utilizing voxel processing to enable real-time morphological and Boolean operations on complex spinal geometries. This specific tool reduced the design time from over one week to under one hour, enabling a massive increase in patient volume.

Medical Image Segmentation (U-Net)

Isolating the exact vertebral anatomy from surrounding tissue was a major bottleneck previously performed manually by radiographers. To solve this, I trained and deployed a custom U-Net convolutional neural network using Python and TensorFlow. The model achieved a Dice coefficient exceeding 95%, matching clinical quality and reducing manual segmentation design time from 1.5 days to just 1 hour per case.

Automated Documentation Generation

Regulatory compliance generates massive amounts of paperwork. I wrote Python automation scripts that eliminated approximately 50% of the manual data entry required for QA and manufacturing documentation, completely removing human error from the regulatory paperwork pipeline.

Infrastructure and Operations

When I joined, 3DMorphic had minimal IT infrastructure. I architected, deployed, and administered the company's entire network and data ecosystem from the ground up:

• Networking and Security: Deployed a Ubiquiti UniFi ecosystem with distinct VLANs isolating production, development, and guest traffic. I implemented a zero-trust network access model using Twingate to safeguard company data from internal and external threats.
• AI/ML Platform: Built and containerized a dedicated onsite GPU server using Docker, providing a high-performance environment for machine learning model training and inference.
• Disaster Recovery: Engineered a robust 3-2-1 backup strategy with automated, encrypted pipelines to both an onsite and an offsite Synology NAS, ensuring business continuity.
• DevOps: Implemented Docker, Gitea, and CI/CD pipelines to enable containerized deployment and strict version control for medical software compliance.

Impact and Results

• Reduced end-to-end implant design time from over a week to under an hour.
• Built the technological foundation that supported Australia's first TGA-approved patient-specific implants.
• Boosted company-wide productivity by eliminating network latency and minimizing operational blockers.
• Championed Agile methodologies, which significantly improved overall project velocity and team workflow structure.