The current project aims to design and develop a new solution for bottleneck detection in microservices, considering their inherent dynamism in production environments.

Transferable skills: Microservices, Root Cause and Anomaly Detection, Graph Neural Networks

During my MS.c., I created ML-Adapt, a proactive system that uses machine learning to forecast CPU for auto-scaling microservices. ML-Adapt notably reduced application response time by 20% compared to HPA in best-case scenarios. However, its effectiveness relied heavily on forecast accuracy. My Ph.D. research focuses on enhancing this forecast component. I introduced the Multiple Predictors System (MPS) approach, demonstrating its superior accuracy (35-75% improvement in the best results) in 81.5% of experiments compared to the previous approach.

Transferable skills: Machine Learning, Microservices, Time series forecasting, Auto-scaling, Self-adaptive Systems, Kubernetes, Python, and Java

This project aims to support the development, execution, and monitoring of service compositions. As service compositions are executed in dynamic environments and developed by different programmers, formal verification techniques are used to ensure the expected behaviour is met during runtime.

Transferable skills: SOA, Microservices, Formal description, Self-Adaptive Systems, and Systems modelling

I created a suite of digital games using augmented reality to improve children’s motor coordination. The games required the child to identify markings containing puzzles, leading them to the next mark. As a result, the software aimed to stimulate children’s movement.

Transferable skills: Unity, Android, and Augmented reality