Vantari VR — Co-Founder / CTO

www.vantarivr.com
2017 – 2025 | Fully Remote

Vantari VR is a medical technology startup delivering virtual reality based clinical training and simulation. I co-founded the business, bootstrapped the early technology, and led the engineering organisation through seed funding and into a pre-Series-A stage, with teams distributed across Australia and the United States.

Leadership and Impact

• Built and led the technical side of the business (~15 technical staff out of ~30 staff at peak), established the engineering culture, practices and delivery focus across a remote, multi-timezone team
• Defined core technology strategy and product direction, aligning engineering capability with business needs across hospitals, universities and medical device OEM markets
• Grew delivery capability from two simulations per-annum (2020) to 20+ simulations per-annum across multiple medical domains (critical care, cardiology, ultrasound) for customers globally (Australia, US, Europe)
• Supported the company’s entry into the US market in 2022, where the platform won key customers in competitive evaluation based largely on technical merit
• Prioritised professional development, mentoring and growth; used design and code reviews as coaching opportunities to build engineering capability across the team

Product Design and Strategy

Contributed significantly to product strategy, aligning market opportunities, technical strategy, and engineering capability. Key fundamental contributions include:

• Human Factors and Product Philosophy
• Championed a Human Factors driven approach to VR training, ensuring interaction models, UI and learning experience design were purposeful, intuitive and outcome-focused
• Established product principles that prioritised training effectiveness over spectacle, helping differentiate the platform against animation-heavy competitors

• Real-Time Simulation Strategy
• Promoted the use of real-time simulation over animation, enabling user freedom, exploration, realistic interactions, and consequence-based learning
• Maintained a pragmatic balance in engineering discipline by defining where to invest in simulation to maximise value without unnecessary risk and complexity
• Delivered a unified, reusable simulation capability to efficiently support diverse medical domains while improving scalability and ROI of technology investment

• Platform and Ecosystem
• Technology stack selection, including Unreal Engine, C++ and PhysX; deployed as a desktop application with supported VR headsets
• Learning Management System built as a multi-tenant web platform (NodeJS API, Vue frontend, Couchbase), hosted on AWS across structured dev/test/prod environments
• Ensured the combined platform was scalable, performant and commercially robust for real-world clinical training environments

Technical Contributions

I contributed heavily across all Vantari products, from desktop VR to full-stack applications. These contributions span multiple platforms, architectures, programming languages and user interaction modalities. To highlight my unique technical contribution to Vantari's core IP, I will provide two examples:

• Generative Real-Time Ultrasound Simulation: Designed and implemented a physically-based generative ultrasound simulation using GPU-accelerated ray tracing via NVIDIA OptiX. The system generated ultrasound images dynamically from anatomical surface models. The simulation naturally modelled acoustic reflections, transmission losses, acoustic shadowing and enhancement, and doppler modes. Tissue texture was simulated by locally evaluated noise models, and beam width via oversampling.

  The system was applied to several contexts, including full abdominal and third-trimester pregnancy models; large, complex geometries that required careful handling of degenerate intersection cases and modelling deficiencies such as interpenetrating surfaces. The resulting image fidelity was clinically compelling and significantly differentiated the platform in competitive evaluations.

• XPBD Soft-Body Physics Solver: Designed and implemented a custom XPBD (Extended Position Based Dynamics) physics solver to support real-time soft-tissue interaction, including incisions, needle insertion, and ultrasound probe pressure deformation.

  XPBD was chosen specifically for its stability under strict time-budgeted simulation - a critical property for interactive VR simulation where physical plausibility and stability matter more than physical accuracy. This simulation approach enabled clinically meaningful deformation responses on surface and volumetric datasets.

• Parallel Processing and Resource Balancing: Architecture, design and implementation of systems for parallel processing and resource balancing within simulations, promoting graceful degradation under resource pressure to maintain interactivity.

• VR-Embedded UI System: Designed and implemented a flexible, modern UI system for VR-embedded user interfaces, based on Vue and rendered by an embedded HTML renderer.

• Platform and Tooling: Established tech stack (Unreal Engine, C++, NodeJS, Vue, Couchbase), build systems, CI/CD pipelines, and dev/test/prod environments (AWS).

Performance Does Matter — Principal Consultant / Engineer

2016 – Present

Independent practice, created as my personal vehicle for consultation, built on a conviction that software performance and design discipline still matter in an age of abundant compute resources. Technology-agnostic and outcome-focused, engaged across varying technology stacks, both legacy and modern. Emphasis on solving real problems and delivering tangible, measurable results. Operated as a selective, high-impact solo consultancy.

Selected Engagements

• Mining automation rescue (major global mining company): Engaged to assess a decade-long, failing automation programme (>100 person-years) and determine salvageability. Recommended a full architecture and stack reset with a 4–6 person-year delivery plan for MVP; assembled the core technical team, delivered MVP, and successfully transitioned leadership to an in-house team for continued development.

• AI surveillance platform stabilisation: Conducted a short two-man-week technical assessment of an early-stage AI surveillance platform struggling with various technical issues; identified a critical defect that had eluded the team via static analysis, and delivered an architectural roadmap that continued to guide their engineering practice and trajectory over the following year.

• Database scaling, mining software provider: Worked with a mining software provider to resolve database scaling issues, designing a custom bitmap indexing strategy and consistency model that enabled performant, predictable queries.

Think Bottom Up — Co-Founder / Consultant / Engineer

2008 – 2016

Created this consultancy to counter observations of top-down decision-making and implementation eroding technical value and creating inertia in complex systems. Focused on grounding solutions in strong fundamentals, practical engineering discipline and bottom-up technical insight.

Selected Highlights

• 10,000× GPU query engine — mining resource model visualisation (major mining company): The client needed real-time interactive queries against an in-house resource model management system whose architecture made this fundamentally limited and could not be changed within the scope of my project. Proposed and led the design and implementation of a custom columnar store built on a log-structured, layout-conscious data layer, with a GPU query layer executing dynamically compiled OpenCL kernels orchestrated across multiple GPUs with careful local memory management for high occupancy. Baseline query time for a visualisation slice was minutes; the new system responded cold in under a second, and hot in a few milliseconds — a realised improvement of approximately 10,000×. Served as architect, design lead and OpenCL technical lead throughout; grew and led the team to seven engineers over several years.

• HPC optimisation — seismic data processing, oil & gas: Implementation and optimisation of core HPC codes for a seismic data processing provider, including Kirchhoff Migration and SRME, across a heterogeneous CPU/GPU cluster (the largest private compute cluster in the southern hemisphere at the time), where cluster occupancy was a key commercial metric. Delivered significant performance improvements, including approximately 10× on SRME, in workloads where 5% gains were considered commercially impactful.

• Distributed biometric voter registration platform: Designed and delivered a distributed biometric voter registration platform used in developing regions, including hardware design, software architecture and distributed database synchronisation; enabled reliable offline operation with robust two-way data reconciliation to a central system, significantly improving deployment reliability and data integrity in remote and offline environments.

Fractal Technologies - Technical Director (multiple roles)

Graduate Engineer → Senior Engineer → Development Lead → Technical Director
1998 – 2008

Joined straight out of university into Fractal Graphics (which later became Fractal Technologies) and grew through senior engineering and leadership roles, ultimately becoming Technical Director and later Software Architect briefly after acquisition. Worked in a uniquely empowering engineering culture focused on technical excellence, close collaboration with domain experts, and practical delivery.

• Instrumental in the architecture, design and delivery of Fractal’s core products, including advanced mining data management and 3D visualisation systems used across exploration and mining
• Played a key role in commercialising research, including work with CSIRO (haptic workbench, seismic inversion, geological data modelling), and in evolving the technology from specialised SGI workstation environments to capable desktop platforms as GPU hardware matured
• Led the architecture and early delivery of a new enterprise data management platform with workflow, compliance and audit capability; the product’s success led to competitive displacement of larger vendors and ultimately to Runge acquiring the business
• Continued post-acquisition with Runge for a short time as Software Architect to support transition, alignment and knowledge transfer into Runge’s product ecosystem

Astrometrical — Founder / Sole Engineer

www.astrometrical.com
2025 – Present

Astrometrical is a platform for publishing and exploring astrophotography, with a focus on public outreach and education. A full-stack solo project, spanning product vision, UX/UI design, architecture, and implementation; a rare opportunity to exercise all capabilities in combination.

• Designed and implemented high-performance image handling for very large astronomical images (eg, 5 gigapixel outputs from Vera C. Rubin Observatory); client-side processing in the browser handles the heavy lifting, keeping infrastructure lean while supporting uploads of 25GB+ TIFF files
• Implementing an AI-guided virtual tour experience: a context-aware agent that leads users through astrophotographs, adapting explanation depth and tone to the audience (eg, a classroom of primary-school students vs. an advanced amateur)
• Applying and refining agentic AI-assisted development workflows throughout, treating the project as a testbed for modern engineering practice as much as a product

Astrophotography Image Restoration - Software Developer

2025 – Present

Independent implementation of the ImageMM algorithm, published by researchers associated with the Vera C. Rubin Observatory. Working implementation producing good results; ongoing work addressing ringing artefacts around bright stars. Developed entirely through agentic AI-assisted workflows. Demonstrates the recurring pattern of taking published research and turning it into working, practical code.

Bachelor of Engineering (Honours)
University of Western Australia

• Astrophotography, astronomy and related science
• Engineering, prototyping, 3D printing and practical making
• Audio, music and immersive listening experiences
• Technology, innovation and exploration of emerging tools and platforms