Speakers · DevFest Skopje 2026

Meet the speakers.

Nine speakers, nine talks, one day on Building with AI Agents. Here is who is on stage and what they will cover.

10:45 – 11:30 · 28 November

A Tour of Agent Protocols: MCP, A2A, A2UI, and AG-UI with ADK

We are rapidly moving from a world of isolated AI tools to an ecosystem of interconnected agents. To survive this shift, developers need to master the protocols that allow these systems to communicate. This session takes you on a guided tour of the most critical agent protocols: MCP (Model Context Protocol) for secure tool access, A2A for agent-to-agent negotiation, and A2UI/AG-UI for agent-generated frontends. Through live, hands-on demos using the Agent Development Kit (ADK), we will wire these protocols together to show how specialized agents can seamlessly pass context, delegate tasks, and render bespoke UI components on the fly.

Attendees will move past building isolated AI scripts and learn the architectural standards required to build interoperable agents that seamlessly plug into enterprise ecosystems, other AI models, and modern frontend frameworks.

Level: Beginner, Intermediate · Speaker based in London, United Kingdom

Vasudev MaduriStaff Data Engineer, GDE, Mentor, Cloud EvangelistAdmiral Group Plc

About Vasudev

Vasudev is a seasoned professional currently serving as a Principal Engineer at ISS Stoxx (Deutsche Börse Group). With a wealth of experience in Google Cloud Platform (GCP), big data, and data architectures, Vasudev has consistently demonstrated his prowess in implementing industry-leading products recognized by Gartner. His journey in the world of technology has encompassed stints at organizations like Hitachi and Ford, where he honed his expertise in data engineering and analysis. Vasudev is not only a practitioner but also an active mentor, deeply committed to fostering talent and contributing to open-source technologies. As a thought leader in his field, Vasudev has shared his insights and knowledge with a global audience through over 30 impactful talks, mentorship programs, and supporting startups. Vasudev continues to inspire and innovate, driving transformative change in the world of data and technology.

Hover or tap for full details

11:30 – 12:10 · 28 November

Autonomous Remediation with CodeMender and Google Cloud

Modern software supply chains generate security alerts faster than engineering teams can triage them. Static analysis tools and dependency scanners uncover thousands of vulnerabilities, but remediation remains painfully manual. Simple LLM code generators often fall short: they lack repository context, hallucinate non-existent APIs, and fail to verify whether fixes break adjacent unit tests.

In this practical, code-first session, we explore CodeMender, an agentic remediation system designed to diagnose, rewrite, and verify code at repository scale. We examine how CodeMender decomposes vulnerability repair into an iterative loop: parsing abstract syntax trees (ASTs), formulating semantic patches with Gemini, and deterministically validating fixes using isolated runtimes on Google Cloud. Attendees will discover how to transition from passive alerts to an autonomous remediation pipeline that submits verified, compilable pull requests directly into their CI/CD lifecycle.

Key takeaways

Deconstructing CodeMender: how agentic code-repair loops analyse, patch, and verify repository vulnerabilities.
Mitigating regression risks through automated test execution and semantic diff validation.
Running sandboxed remediation environments using Cloud Build and Cloud Run.
Integrating autonomous remediation agents into existing enterprise CI/CD and workflows.

Level: Intermediate, Advanced · Speaker based in London, United Kingdom

Goran MinovSenior Technical Account ManagerOkta

About Goran

Goran Minov is a Cloud AI & Cyber Security Architect and Google Cloud Certified Professional, holding key Google Cloud credentials, including Professional Cloud Architect, DevOps, and Database Engineer. Specialising in the intersection of applied machine learning and Identity & Access Management (IAM), he currently serves as a Senior Technical Account Manager at Okta, where he helps customers automate and secure their identity infrastructure for modern, intelligent workloads.

A passionate community leader, Goran has been an Organiser for GDG London since 2019, helping grow one of the largest developer communities in the UK. He is also the Founder of The Cloud Circuit, an independent event series connecting London’s practitioners with advanced cloud and AI deployment strategies, and a Co-founder of The Android Circuit.

As a prominent speaker and technical mentor, Goran regularly shares his expertise in securing Agentic AI and implementing Zero Trust Architecture. He has delivered workshops and talks at major venues including Droidcon London and DevFest.

Hover or tap for full details

12:25 – 13:05 · 28 November

Talk title to be announced

Stefan AndonovTeaching Assistant, Data Engineering LeadFCSE

13:05 – 13:45 · 28 November

Talk title to be announced

Teodora MladenovskaSenior ML EngineerPythian

14:45 – 15:25 · 28 November

Fast & Bulletproof: Securing and Scaling the Gemini Live API

Integrating the Gemini Live API for real-time, conversational AI creates incredible user experiences, but introduces complex architectural challenges. When building deutschexam.ai, a real-time language exam prep tool, balancing sub-second latency with strict AI guardrails was our biggest hurdle.

This technical session bridges the gap between performance optimization and AI security. Drawing from real-world implementation, I will demonstrate how to architect secure, continuous AI streams that handle unpredictable user inputs without breaking.

We will cover how to manage live bidirectional connections efficiently, cache conversational context to reduce API overhead, and actively sanitize real-time streams to block prompt injection attacks—all without sacrificing response speed. Attendees will leave with a concrete blueprint for building real-time AI feedback systems that are both blazing fast and secure.

Key takeaways

Defending the Stream: Techniques to detect and block prompt injections during live, continuous AI sessions.
Zero-Latency Caching: Strategies to cache contextual data and minimize API round-trips.
Memory Management: Keeping client and server overhead low in persistent, real-time architectures.
Production Guardrails: Simple, actionable steps to protect user data and ensure predictable AI behavior.

Level: Beginner, Intermediate, Advanced · Speaker based in Mainz, Germany

Balram ChavanPrincipal Software Engineer, Google Developer ExpertLucaNet

About Balram

Principal Software Engineer | Cloud Architect | Google Developer Expert Angular | Author

Hover or tap for full details

15:25 – 16:05 · 28 November

Interactions that work for all users

This talk explores how users interact with websites and how developers can make those interactions accessible to everyone. Using practical examples, such as converting an inaccessible navigation menu or carousel into accessible versions, we’ll explore accessibility issues with complex components and their solutions. The session will include multiple coding demonstrations to illustrate how intentional code changes can significantly impact digital inclusion.

Session takeaway: The audience will understand how users can use the websites in different ways, they will learn different ways to test a scenario, and practical code solutions on how to make the interactions accessible.

Level: Intermediate, Advanced · Speaker based in Haarlem, Netherlands

Anuradha KumariSenior Front-end DeveloperGoogle Developer Expert

About Anuradha

Anuradha Kumari is a senior frontend developer from India, currently based in the Netherlands. She is passionate about exploring new technologies and sharing knowledge through tech articles and talks. As an accessibility advocate, she aims to spread awareness and empower the community towards achieving the common goal of inclusion through technology. When she is not coding, she loves traveling, exploring museums, reading books, and sketching.

Hover or tap for full details

16:25 – 17:05 · 28 November

Architecting Scalable MLOps Solutions on Google Kubernetes Engine

We will dive into the architecture of an AI/ML platform built on GKE, highlighting the importance of portability, flexibility, scalability, and cost efficiency. The session will also cover the latest advancements in GPU and TPU integration, and how these technologies can accelerate AI workloads. Real-world use cases, ML training and ML serving, will illustrate the practical applications and benefits of GKE in managing AI/ML workloads at scale.

Level: Intermediate · Speaker based in Istanbul, Türkiye

Enes TuranPlatform Software EngineerJumpCloud

About Enes

Enes is a Google Certified Professional Cloud Architect who specializes in cloud and platform engineering, Kubernetes, Google Cloud, Anthos, and Apigee consulting in the region with a Google Cloud Premier Partnership and Application Modernization Specialization. He has assisted multiple organizations in implementing DevOps practices and improving the reliability and scalability of their applications on Kubernetes. He also has experience in designing and implementing cloud infrastructure, and a deep understanding of best practices for cloud security, compliance, and governance. Enes has a strong passion for sharing his knowledge with others and helping them to be successful with DevOps, Kubernetes, and Cloud Architecture. He is organizer at Cloud Native Istanbul, instructor at Udemy and EC-Council.

Hover or tap for full details

17:05 – 17:45 · 28 November

Talk title to be announced

Nino RisteskiML & Data LeadIVTM Analytics

17:45 – 18:25 · 28 November

AI in Production: Sampling, Labeling and Detecting Drift Before Your Customers Do

AI systems don't always fail when they go into production—they drift. Documents change, data distributions shift, vendors introduce new formats, and models behave differently on real-world inputs than they did in testing.

In this talk, we’ll build a practical evaluation loop on Google Cloud around a real-world AI document-processing pipeline. We’ll explore how to sample production data, create reliable ground truth through human labelling, evaluation, and detecting drift before it becomes a customer-facing problem.

We’ll walk through a live demo with Vertex AI, Document AI, BigQuery, Cloud Storage, and Cloud Run to create a continuous evaluation and feedback loop. The goal isn’t another accuracy dashboard. It’s a production feedback loop that tells you when your AI system is changing, why it changed, and what to do about it—before your customers do.

Level: Intermediate, Advanced · Speaker based in Berlin, Germany

Akshata MohantyGoogle Developer Expert in AI/ML, Head of EngineeringTilla

About Akshata

Akshata Mohanty is a Google Developer Expert in AI/ML and Head of Engineering at Tilla Technologies, Berlin. She founded Keis, a Singapore-based consultancy creating generative AI products for startups and venture studios. She has held senior roles at Rakuten, Nas.io, and the National University of Singapore, and speaks internationally on generative AI, MLOps, and cloud architecture.

Hover or tap for full details