I am an inquisitive and energetic team leader that enjoys deep dives into data and real-world problems with experience leading projects from inception to commercial deployment. Coming from a startup background, I have a versatile set of skills, where I can come in develop on Android, all the while building the machine learning models that power the app. At that core I am an applied engineer that loves to build and tinker.
Developed physical exercise recognizer and repetition counter including data collection, neural network development, model training, and real-time iOS implementation with off by 1 error of 1%.
Led and implemented tech transfer with external vendor for 2 month timeline on custom Samsung hardware running Android for exercise recognition.
Collaborated with Samsung AI Centers to productionize their R&D technologies for facial analysis.
Built reproducible ML workflow including versioning data, training runs, results, and models.
Built an end-to-end deep facial analytics (age, gender, and identification) system powering all products including in-car identification of <1% false accept and reject rates.
Deployed deep license plate recognition system improving legacy system by 20%.
Sped up in-house library with CUDA implementations of convolutional depth-wise layers.
Thesis: A Scalable and Efficient Outlier Detection Strategy for Categorical Data
Advisor: Professor Michael Georgiopoulos, Dean and Director of Machine Learning Lab at UCF
