Projects

Kara Technologies

(As of early January 2022) During my time at Kara, my co-founders and I have so far raised upwards of $1.5mil in investment capital; and have developed what is acknowledged by tech giants such as Cisco and Microsoft and international leaders from the Deaf community as the most advanced virtual sign language platform.

We have established deep relationships with the Deaf community that inform our product development. Hand in hand with this we have built significant strategic relationships with two world leading organisations Epic Games and Vicon to enable our solution. These companies have in turn invested funding, technology, and other resources to support Kara achieve our goals.

We have generated pilot sales of over $200k and have a near term opportunity pipeline of a further $200k. We have gained international traction and are on a path to a $1mil+ ARR in 2 years’ time.

While playing a part in the achievements of the Kara team, I am personally in charge of the internal technology development. On a day-to-day basis my responsibilities range across:

  • Managing company’s technology plan, including the scoping and execution of said plans.
  • Envisioning new forms of product/services that will best strengthen the business.
  • Together with the CEO, designing the full architecture of a unique and first-in-the-world software platform for providing sign language accessibility.
  • Responsible for 70% of company’s cashflows over the past 4 years, mainly on R&D and COGS.
  • In charge of full product life cycle, including directly managing up to 4 engineers at any one time.
  • Design and development of systems architectures and high-level UML diagrams, utilizing optimal design patterns and architectural guidelines.
  • Simultaneously responsible for the delivery of all services with a team of up to 9 animators and sign language experts.

My engineering and programming involvements have been quite varied across different fields. My main focus has been on C++, Unreal Engine, and Python scripting. A few related examples are as follows:

  • A quaternion-based real-time motion transition algorithm for the generation of transition graphs between 2 pieces of animations. The algorithm is context-sensitive and can generate varying animations according to the context and field of motion. This technology is currently patent pending.
  • Algorithms for automatic viseme generation on the avatars
  • A high-quality real-time render pipeline for Unreal Engine to render MetaHuman avatars and our custom avatars. This involved complete project workflow including set design, lighting, blueprinting, Python-based automations, and C++ algorithms.
  • Internal tools built in Unreal Engine for the review and processing of avatar animation data. Required development of communication mechanisms, and database controls.
  • A custom python-based command centre with user GUI, for control management and recording of Motion Capture data.
  • Machine Learning R&D experiments for product improvements, conducted in Python.