Eduard Kieser Resume

Eduard Kieser - Résumé 2023

Personal Information

  • Location: Cape Town
  • Date of Birth: 18 December 1990
  • Phone Number: +27 73 218 5933
  • Email: eduard.kieser@gmail.com
  • Website: eduardkieser.github.io
  • Languages: English-Native, Afrikaans-Native, German-Proficient
  • Licenses: Driver’s (B), RPAS (Fixed Wing and Multi-Rotor Drone Pilot License)

PROFILE

I am a competent and motivated Engineer, Data Scientist, and Design Thinker. I have over 8 years of experience in applying technology to solve complex problems in healthcare, automation, and aviation. I enjoy working in a team and I learn quickly.

EDUCATION

  • 2013 - 2015: Master of Engineering (MEng, Mechatronics), Stellenbosch University, South Africa. (Cum-Laude)
  • 2009 - 2012: Bachelor of Engineering (BEng, Mechatronics), Stellenbosch University, South Africa.

WORK EXPERIENCE

  • Oct 2023 - present: ML Engineer and System Architect at BlueCube Systems.
  • Refined models and model training infrastructure for sensors used in industrial process control (lasers and spectrograms).
  • Developing cloud architecture for next generation system deployment.
  • April 2020 - August 2023: Machine Learning and System Design Engineer at 42air.
  • Developed ML models and embedded systems for delivery drones. Most of the ML work here involved modifying models and setting up training infrastructure to retrain modified SOTA vision models. The tasks included segmentation, pose estimation, depth estimation and object detection. All of which had to work in real time within the hardware constraints of our embedded system. There was also a good amount of 'classical' CV work.
  • I was also responsible for our in-house vehicle development program. This included everything from concept to test flying. By the end we had test flown three VTOL prototypes, the largest of which had a 3m wingspan, a 5kg payload capacity and 25kg MTOW.
  • Aug 2018 - March 2020: Data Scientist at HealthQ, Stellenbosch.
  • Guided technical development and refined algorithms for wearable fitness devices. I played a key role in transforming skilled scientists into efficient programmers. Much of my role involved sustaining production models by monitoring drift and integrating the latest scientific advancements into our products. At healthQ, I discovered the thrill of adapting machine learning architectures to achieve previously unattainable results. My work also ventured into self-supervised learning and complex/cascading models. Two notable research projects included identifying arrhythmias from noisy wearable data and detecting sleep apnea using smartwatch sensors.
  • Nov 2017 - Aug 2018: Consulting Data Scientist at Neural Sense, Cape Town.
  • Developed Memory Salience assessment platform and integrated it into existing services.
  • Nov 2015 - July 2016: Design Thinking Coach at UCT d.school, Cape Town.
  • Founding member, also coached at Hasso Plattner Institute in Berlin.
  • Mar 2014 - Aug 2018: Program Coordinator for the Global Education Teams program.
  • Facilitated communication and advised students on projects.
  • Dec 2015 - Aug 2018: Founder of BeatCounter.
  • Your medical data should not be thrown away or locked away in some OEM's basement. This was the founding principle of beat counter. To meet this goal I developed a completely general medical device data-logger. The only shared interface between all these devices is of course the screen, so it was a vision based solution. More information on the solution can be found here, the patent mentioned below also relates to this work.
  • Apr 2015 - Oct 2015: Intern at Philips Research, Eindhoven.
  • Continued the development of an algorithm for pulse detection, resulting in an invention disclosure.
  • Dec 2014 - Mar 2015: Researcher at Biomedical Engineering Research Group.
  • Redesigned a wear tester and wrote supporting legal documentation.
  • Nov 2012 - Feb 2013: Intern at Desert-Wolf, Pretoria.
  • Assisted in radar and UAV platform design.

RESEARCH EXPERIENCE

  • 2015 - 2018: Exploring maternal heart rate patterns during pregnancy as a predictor to determine infant prognosis.
  • Research with Professor Hein Odendaal and partners at Tygerberg Hospital and Columbia University.
  • Set up data transformation pipeline and translated clinical ideas into code.
  • 2013 - 2015: Development of a non-invasive infant hydration sensor.
  • Master's thesis project, developed embedded systems to check hydration markers.
  • 2013: Development of a circulation chamber for improved testing of infective endocarditis.
  • Part of Global Engineering Teams program; worked on a project at Tygerberg Hospital.
  • 2012: Mini Autonomous Ground Vehicle Design.
  • Bachelor's thesis, developed an autonomous ground vehicle (AGV).

PUBLICATIONS

PATENT AND INVENTION DISCLOSURE

  • South African Patent Application 2018/04931 (page 264) was filed on my behalf by Stellenbosch University, with myself as the sole inventor. It was granted in April 2019.
  • A invention disclosure was filed for a pulse presence detection algorithm based on accelerometry data that I helped develope at Philips.

PERSONAL PROJECTS

My personal projects are numerous and span hardware, software, aircraft design, robotics, machine learning and art. Some of them are listed on my personal website under Past Projects.

TECHNOLOGIES

  • Scientific Programming: Python, MATLAB, SMath-Studio, MathCad
  • ML and Data Science Tools: PyTorch, TensorFlow, Pandas, SciPy and sklearn.
  • Mobile/Web/General: Dart, Flutter, JavaScript, Go, C/C++
  • CAD: Inventor, Fusion 360, Onshape, Solid Works, Ki-CAD
  • Infrastructure & Deployment: Ansible, AWS EC2, S3, Sagemaker, Lambda, GCP and many more.
  • Miscellaneous: Docker, Bash, Zsh, Git, Markdown and many more.

INTERESTS

Building and flying model aircraft, DIY projects (gadgets and furniture), design, kite-surfing, surfing, rock-climbing, running. Using technology to make a positive impact.

References available on request.