Romualdas Randamanskas

Took part in the PyTorch Scholarship Challenge from Facebook, where only the top 300 students from the initial Challenge Course were selected for the Deep Learning Nanodegree program.

Syllabus:
• Neural Networks
• Convolutional Neural Networks
• Recurrent Neural Networks
• Generative Adversarial Networks
• Model deployment with AWS SageMaker.

The Facebook Deep Learning Scholarship program Phase-3 also awarded 200 scholarships to the Deep Reinforcement Learning Nanodegree program in which I participated.

Syllabus:
Foundations of Reinforcement Learning
• Value-Based Methods
• Policy-Based Methods
• Multi-Agent Reinforcement Learning

In this program, as a student, I've learned how to run and manage scalable applications in a cloud-native environment using open source tools and projects like ArgoCD, gRPC, and Grafana. And I've also learned how to identify the best application architecture solutions for an organization's needs, design a microservice architecture by leveraging cloud-native tools and patterns, implement best practices in Kubernetes security, and use dashboards to diagnose, troubleshoot, and improve site reliability.

Syllabus:
• Cloud Native Fundamentals
• Message Passing
• Observability
• Microservices Security

Participated in the NASA International Space Apps Challenge as part of the EQsight team project. The team that worked on the update of QuakeHunter, the earthquakes activity analysis app.

The EQsight noticed that valuable earthquake catalog API data was openly provided by USGS and decided to use it to upgrade the existing earth data visualization app QuakeHunter. We used GeoJSON data once it could be used as a point of reference for advanced predictive models, anomaly detection, and distributional analysis. After reviewing the QuakeHunter repository, we focused on refactoring the earthquakes activity time-series graphic, updating the algorithm to correctly count daily earthquakes for a given coordinate point and radius. Additionally, we added a feature to find the maximum daily magnitude on the Richter scale to test the hypothesis that an increased number of earthquakes could distinguish higher magnitude earthquakes.