Konstantinos Lianos
Cloud Security Specialist - Microsoft Regional Leader & Senior Microsoft Student Ambassador
Athens, Greece
Actions
Cybersecurity professional specializing in Research & Development, Multi-Cloud Security, and SecDevOps, currently working across two complementary roles as a Senior SecDevOps Engineer and Cloud Security Specialist.
My expertise spans Microsoft Azure, AWS, Oracle Cloud Infrastructure (OCI), and Google Cloud Platform (GCP), with a strong focus on cloud security architecture, SIEM/XDR, threat detection and response, IAM, security automation, DevSecOps, and cloud-native defense.
With an MBA, MSc in Information Security & Digital Forensics, and BSc in Computer Science, I combine technical depth, security research, and business understanding to design and improve secure, scalable multi-cloud environments and translate complex cybersecurity challenges into effective security solutions.
Area of Expertise
Topics
See Your Attack Surface: Mapping Azure with an MCP Server
Most breaches start with something nobody knew was exposed. If you cannot see your environment clearly, you cannot defend it.
In this session I build an MCP (Model Context Protocol) server that scans an Azure subscription and renders it as an interactive graph, then I turn that graph into a security tool. We look at how resources actually connect, where the exposed edges are, and how to reason about blast radius when one resource is compromised.
I cover the practical safeguards that matter when you point AI at a real tenant. Keeping the scan strictly read-only, handling large subscriptions without noise, and building in leak-detection so sensitive data never lands in a prompt. I also share the precision rule I enforce throughout: every security finding has to come from actual scan data, with gaps marked unknown rather than inferred from a resource name or tag. No guessing your way to a false sense of safety.
I close on where this goes next, using the same map to reason about Defender for Cloud coverage and which logs are actually flowing into Sentinel.
You will leave understanding what MCP is, why it fits security tooling so well, and how to start building an Azure-aware assistant that helps you see what you are actually exposing.
Who it is for: security and cloud engineers who want a clearer, defensible picture of their Azure attack surface.
Turning AI Governance into Controls You Can Prove
Most organizations now have an AI policy. It promises human oversight, data protection, access control, and responsible use. Many are mapping to ISO/IEC 42001, the NIST AI RMF, and the EU AI Act, while also meeting GDPR and NIS2. On paper, the governance is in place. In practice, few of these promises are backed by a technical control, and even fewer by evidence an auditor or regulator would accept.
This session tests an AI agent that passes a typical governance review. Every policy box is ticked. Then it gets hijacked through a poisoned document and leaks data, and none of the stated controls stop it or even record it.
We then work in both directions. Upward, we turn the technical failure into GRC terms: the risk register entry, the control that failed, the regulatory exposure, and the questions auditors and regulators will ask. Downward, we take common AI governance requirements such as human oversight, logging, access limits, and data protection, and turn each one into a specific technical control, an evidence source, and a way to test it.
Hijacked by a Document: Catching a Rogue AI Agent in Your Logs
AI agents now read email, open files, and call cloud APIs on our behalf. One poisoned document is enough to turn that agent against you. It uses the agent's own access and valid credentials, and nothing in your stack sees malware. To the SOC, it looks like normal work.
In this session, I hijack a working AI agent in a lab using indirect prompt injection and follow the attack from start to finish through the logs. You will see what the agent did, what evidence it left behind, and which signals give it away: reading data it never touched before, calling tools in an unusual order, sending data to new destinations, and working at a speed no human could match.
From that trail, we build detections step by step using identity, cloud audit, API, and network logs that most organizations already collect. We then cover how to baseline a non-human identity, how to tell a hijacked agent from a buggy one, and how to shut it down without breaking the business process it supports.
Original lab research. No products, no hype. You leave with detection logic you can build the next day.
Learning objectives:
See how a hijacked AI agent behaves and why traditional detections miss it.
Trace an agent attack end to end through common log sources.
Build behavioral detections and baselines for non-human identities.
Contain a compromised agent with a clear response playbook.
Microsoft Security Copilot as an Integration Layer for Your Security Estate
In this session we look at Microsoft Security Copilot not as a chatbot, but as an integration layer across the SOC, the thing that ties together SIEM, XDR, identity, threat intel and ticketing instead of leaving the analyst to pivot between them. We walk through how it works, then go hands-on: building custom plugins (including a real KQL plugin over Sentinel), orchestrating with Logic Apps and promptbooks, grounding Security Copilot on your own SOPs, and deploying autonomous agents you can govern with Agent 365. We see it all come together on a single phishing incident handled end-to-end, and close with the SCU cost model and a practical rollout plan you can take back to your own environment.
Building Microsoft Sentinel Codeless Connectors Faster with an AI-Assisted
I will demonstrate how an AI-assisted skill and knowledge base can accelerate the development of Microsoft Sentinel custom data connectors using the Codeless Connector Framework and Solution V3 structure.
The session will walk through the practical workflow of turning a vendor API into a deployable Microsoft Sentinel connector: selecting the correct authentication method, understanding API response envelopes, handling pagination, defining streams, writing supported KQL transformations, wiring DCR stream names, building multi-connection support, packaging the solution, and validating it with ARM-TTK.
I will also explain what cannot be fully automated and why live validation remains critical. This includes collecting real API sample responses, testing paging behavior, validating incremental time cursors, deploying to a test workspace, and confirming that logs are actually ingested.
Freestyle Summer Jam 2026 User group Sessionize Event
Please note that Sessionize is not responsible for the accuracy or validity of the data provided by speakers. If you suspect this profile to be fake or spam, please let us know.
Jump to top