costopoulos

Designed architectures & solved optimization problems for OpCos worldwide using AWS, Azure, GCP and code. Cut $29,361.80 monthly in AWS by detecting & resolving anomalies.

Among 2% (1000+ applicants) of the Vulcanus in Japan 2022-2023 Scholarship, and 1 of the 2 selected by NTT. Spearheaded research & solo development of bleeding-edge FaaS-based Cloud-to-Edge platforms, boosting development speed by ∼ 60%; integrated Kafka, Redis, MQTT for prototyping. Led business research, authored market report, and designed scalable cloud architecture & APIs with best practices; secured executive approval and funding.

Two research projects on HJ Reachability using RL under Professor K.G. Vamvoudakis. First Project: quadrotors and lunar landers, new method of reachability decomposition with RL (46-61% speed improvement). Second Project: devised a Q-learning algorithm for calculating reachable sets. Received A+ mark, very high output of research, and wrote the manuscript and Master’s Thesis off of it.

First spot out of 255 teams participating from all over Europe in the 3rd CASSINI Hackathon: (Re)Visit Europe. The idea was praised by the jury as the best technical implementation of recent years. Trail safety evaluation, cartography through cutting-edge computer vision techniques.

Implementing detection algorithms on both CPU and GPU to track multiple weeds in a field under various conditions. Stereoscopy and stereo camera-calibration using OpenCV’s C++ libraries. Decreased error to 1%. Set up a CI pipeline for automating all unit tests using bash scripts and GitHub Actions.

Two-person end-to-end (from design to deployment) fullstack development of a delivery app using Django, Flutter, VueJS and RESTful API protocol. Composing scalable web and mobile apps and creating administrative platforms with innovative observability tools.