Periklis Korontzis - Mobile Designer | ContraWork by Periklis Korontzis
Periklis Korontzis

Periklis Korontzis

Computer Engineer focused on machine learning and big data.

New to Contra

Periklis is building their profile!

Followed by GALLERY L

Patras, Greece

WorkAbout

Patras, Greece

Cover image for Developed a comprehensive mobile application
Developed a comprehensive mobile application designed to optimize strength training. The app features a modern, dark-themed user interface and enables users to track live workout sets, detect performance plateaus, and monitor muscle group recovery percentages. Engineered using Flutter for a responsive cross-platform experience, with a local SQLite database ensuring robust, offline-first data persistence and high-speed data retrieval.
1
16
Cover image for Description: Designed and deployed an end-to-end
Description: Designed and deployed an end-to-end real-time big data pipeline to process live vehicle positioning data. The architecture leverages Apache Spark Structured Streaming and Redpanda for high-throughput data ingestion and stream processing. Processed data is seamlessly stored in MongoDB and served to client applications via a custom, high-performance RESTful API built with FastAPI. The entire microservices environment was containerized and orchestrated using Docker Compose for reliable and scalable deployment. Tools & Technologies (Tags): Apache Spark, Redpanda (Kafka), MongoDB, FastAPI, Python, Docker Compose, Data Engineering, Stream Processing.
0
12
Cover image for Designed and developed an end-to-end
Designed and developed an end-to-end deep learning system for the automated analysis of red blood cells from image data. The primary objective was the accurate estimation of critical biophysical parameters, such as cell diameter and minimum/maximum thickness, based on complex morphological features. Architecting and training custom Convolutional Neural Networks (CNNs) and Denoising Autoencoders to process grayscale image patches and extract structural data. Implementing Explainable AI (XAI) techniques, including Class Activation Mapping (CAM) and heatmaps, to interpret model decisions and verify that the network focused on the correct geometrical boundaries. Bridging the gap between computer vision and medical imaging by providing robust, data-driven parameter estimation. Tech Stack & Tools: Python, PyTorch, Deep Learning, Computer Vision, Explainable AI (XAI), Docker, Linux
0
27