SysOps/DevOps Gdańsk MeetUp #31

Welcome to SysOps/DevOps Gdańsk 19.11.2026 godz. 18:00 Q&A session after each lecture Hosted by: Małgorzata Świderek Gdańsk, Hapag-Lloyd Poland office, Grunwaldzka 413 avenue, Neon building, 7. Floor Harmonogram: 18:00 Networking 18:15 Lecture #1: “Agent AI with access to everything. What could go wrong? Tomasz ' Harry ' Zalewski 19:20 Lecture #2: “Anomaly detection in difficult data sets” - Jakub Malczewski 20:15 Networking i pizza! Remember to sign up for the local group - here Gdańsk https://www.meetup.com/sysopsgd/ Videos from previous MeetUpów can be found on YouTubie https://www.youtube.com/@SysOpsDevOpsPolska Partnerzy: MDDV http://www.mddv.pl Hapag-Lloyd Polska https://knowledgecenter.pl/ Akamai https://jobs.akamai.com/en/sites/CX_1/ IBM https://meetup.cloudpak.site/hashicorp-vault/ Lecture #1: “Agent AI with access to everything. What could go wrong? Tomasz ' Harry ' Zalewski Is it possible to create an AI agent, manipulate it, and then protect against manipulation using the HashiCorp Vault? And all this in 20 minutes? During the speech we will show that it is possible! Tomek will create a simple AI agent to support the HR system, try to force it to do undesirable actions, and then install and configure the HashiCorp Vault. He will integrate an agent with it so that it works safely! No theory and no marketing slides. A real agent, a real Vault, a real problem and a real solution. Tomasz 'Harry' Zalewski - on a daily basis, he helps clients learn about security in the world of cloud, Kubernetes, AI and everything that can spoil the administrator's peaceful sleep. At IBM, he deals with security solutions, but he likes to show them in practice preferably live, with a bit of controlled chaos and concrete examples from the life of DevSecOps. An enthusiast of stand-up and talk about how to do something faster, simpler and safer. Lecture #2: “Detecting anomalies in difficult data sets” - Jakub Malczewski Are Data Science skills useful in SRE work? Oh, yeah! Let’s talk about anomaly detection and correlation detection in difficult data sets. How to handle when a static threshold is not enough for effective monitoring. When ordinary SLO fails and the system is full of nuances and dependencies, we need a new approach. Jakub will talk about how Data Science methods collide with the brutal reality of noisy production data, where nothing is obvious. Jakub Malczewski - Senior II Site Reliability Engineer at Akamai Technologies. Doctor of Experimental Particle Physics and developer of ML/AI systems. At Akamai, it ensures stability of inference of large language models and analyzes metrics from hundreds of thousands of servers. Regulations SO/DO MeetUpów desktop: https://bit.ly/RegulaminSODO