•Responsible for algorithm design at ArcSecond. Developing (using Python, Matlab, and C/C++) and testing multi-sensor fusion algorithms powering the company’s motion tracking applications.
•Designed algorithms used on several hardware platforms from the ground up by collecting and preprocessing raw sensor data in an embedded environment, training and tuning hyperparameters in the cloud, and running models on iOS devices.
•Developed IMU (inertial measurement unit) and AHRS (attitude and heading reference system) algorithms utilizing tools such as deep learning (DNN - deep neural networks, CNN - convolutional neural networks, RNN - recurrent neural networks, LSTM - long short-term memory), conventional machine learning (logistic regression, decision trees, random forests, boosting, bagging, SVM - support vector machine), Kalman-filter and other DSP (digital signal processing) tools.
•Designed the sensor (gyroscope, accelerometer, and magnetometer) calibration protocol, including the in-factory tests.
•Involved in the creation of an in-house visual-based advanced motion capture laboratory.
•Had actively participated in company research related to sensor fusion, deep and machine learning, and motion tracking with an interest in filing IP rights for the aforementioned research topics.