Arkar Min Aung

Deep Learning Engineer

Fulbright Scholar

Education

Master of Science in Computer Science

Worcester Polytechnic Institute
2016 - 2018

Bachelor of Engineering (Information Technology)

West Yangon Technological University
2008 - 2013

Profile

Deep Learning Engineer creating solutions to solve challenging problems in training and testing visual perception systems of self-driving cars.

I have 3 years of working experience at two startups in Myanmar and Thailand. I played a crucial role in developing mobile apps and backend, creating design sprints, doing user experience research and managing customer expectations at both startups.

I did my Master's at WPI as a Fulbright scholar and my research was on student engagement level recognition using probabilistic labels for Machine Learning models, automatic eye-gaze following from static images using end-to-end deep learning and building robust neural networks which are more resilient against adversarial attacks.

Research Interests

● Student engagement level recognition
● Automatic eye-gaze following
● Adversarial attack and defense against adversarial attacks on neural networks

Previous Research

Worked with Professor Jacob Whitehill on building end-to-end deep learning systems to recognize eye-gazes of students and teachers to identify where they are looking at in classrooms. [Paper]

Worked on exploring the effects of harnessing inherent uncertainty in crowdsourced annotations on machine learning algorithms. [Paper]

Publications

Aung, A. M., Ramakrishnan, A., & Whitehill, J. R. Who are they looking at? Automatic Eye Gaze Following for Classroom Observation Video Analysis. Proceedings of the 11th International Conference on Educational Data Mining, 2018. [PDF]

Aung, A., and Whitehill, J. "Harnessing Label Uncertainty to Improve Modeling: An Application to Student Engagement Recognition". IEEE Face & Gesture Recognition (FG), 2018. [PDF]. [Supplementary Materials].

Aung, A. M., Fadila, Y., Gondokaryono, R., & Gonzalez, L. (2017). Building Robust Deep Neural Networks for Road Sign Detection. arXiv preprint arXiv:1712.09327.

Master's Thesis

Aung, Arkar Min. Automatic Eye-Gaze Following from 2-D Static Images: Application to Classroom Observation Video Analysis. Diss. WORCESTER POLYTECHNIC INSTITUTE, 2018. [PDF]

Experiences

Deep Learning Engineer

2018 - Present
Octant Technologies, San Francisco and Yangon

Building end-to-end semantic driving data understanding system which will be used to train and test visual perception systems of self-driving cars.

Backend Engineer

2015 - 2016
Drive Design Studio, Bangkok and Yangon

One of the founding team members of DRVR. Responsible for maintaining, optimizing and expanding device message parsing layer of DRVR backend engine. Played a crucial role in orchestrating the whole process of getting DRVR the first vehicle on DRVR portal. Took multiple responsibilities such as:
● Supervising device installations on vehicles
● Configuring and optimizing reporting behaviors on devices
● Training fleet managers
● Conducting UX research by getting feedback and bridging between the dev team and end users
Contributed DRVR in development, product management, operations and customer relations.

Product Manager

2015 March - 2016 August
Nex Co., Ltd, Yangon

Responsible for the product development of 8KET, instant Android app generator for small shop owners. Worked closely with backend and frontend developers and took responsibility as a proxy between customers and tech team. Conducted UX researches and managed the tech team building the product.

iOS Developer

2013 - 2015 March
Nex Co., Ltd, Yangon

Responsible for creating iOS applications with great User Interface and great User Experience. Shipped both client projects and in house products. Worked closely with a team, adopting Agile development methodologies.
Worked with agency side of the startup during the first year. Have experience on developing range of iOS apps. Worked with product development side of the startup afterwards.

Awards and Achievements

Google Hack For Good 2nd Prize - Built a small web app which helps patients and doctors to identify Melanoma from images taken from smartphones. Built a small Firebase web app, cleaned and pre-processed Melanoma skin data from ISIC, trained and fine-tuned a deep learning model (based on InceptionV3).

Fulbright Scholarship - Awarded by the State Department to pursue Master’s Degree in Computer Science at Worcester Polytechnic Institute.

Telenor Youth Summit 2014 Myanmar Representative - Selected as one of the two delegates from Myanmar to attend Telenor Youth Summit which took place at Oslo, Norway from 9th to 11th December, 2014. Selected for the idea of Open Education Platform where the open online coursewares are distributed via mesh network and accessed via a mobile app.

Myanmar First Hackathon 1st Prize Winning Team - Built prototype solution for a problem concerning with pest and plant disease outbreak in farmlands in 48 hours. Divided the team into groups on building application for Android and iOS, building backend and building admin dashboard.