Agata Chudzińska
CTO / AI Solutions Architect at theBlue.ai GmbH
Poznań, Poland
Actions
Agata specializes in the practical application of AI to solve real-world challenges. Currently, she holds the roles of CTO and AI Solutions Architect at the Hamburg-based AI consulting company theBlue.ai. With 10 years of professional experience in AI and data-related fields, Agata has developed AI-based solutions for a diverse range of clients, from emerging startups to global companies.
Badges
Area of Expertise
Topics
From Language to Reality: Why the Future of AI Is Spatial
After years of rapid progress, scaling Large Language Models on internet text is beginning to show diminishing returns. As datasets saturate and performance curves flatten, a growing part of the AI community is shifting its focus toward a broader paradigm - one rooted not in text, but in the structure of the physical world.
In this session, we explore the accelerating momentum behind World Models and Spatial Intelligence. Visionaries such as Fei-Fei Li and Yann LeCun argue that genuine intelligence cannot emerge from language alone - it requires perception, embodiment, and the ability to model how the world evolves. Instead of predicting the next token, future AI systems will predict the next state of the world - learning through interaction, simulation, and physical reasoning.
Attendees will learn why the next wave of breakthroughs is likely to come not from bigger LLMs, but from models that understand space, dynamics, and cause-and-effect. We will demystify the core ideas behind World Models, examine the current state of Spatial Intelligence, and explore what this shift means for developers, researchers, and the future of intelligent systems.
Large Multimodal Models: What's going on in Computer Vision?
How do computers interpret and understand visual data? This session explores the mechanisms behind computer vision AI models, as well as the advancements in Large Multimodal Models (LMMs) that integrate computer vision and language understanding. We'll dissect the mechanisms that enable machines to process and interpret images, discuss the current state-of-the-art models, and evaluate whether leveraging existing LMMs or training the own AI models is more effective for specific vision tasks. Through analysis of various use cases, attendees will gain a comprehensive understanding of the capabilities and limitations of current technologies in the area of computer vision.
Small models, big impact - and no chatbot required
Not every app needs a multi-billion-parameter AI model. In fact, not every company needs a chatbot. But when was the last time you saw an AI talk that wasn’t about GPT?
This session is a reality check and a practical guide for software developers who want to build AI-powered features that address real business problems without relying on huge, closed models behind an API. We’ll explore how small, specialized AI models can outperform their giant cousins in real-world use cases, especially when speed, privacy, cost, or edge deployment matter.
I'll share lessons learned from years of deploying AI in production and walk through examples where lightweight models delivered powerful results. You'll also learn about open-source tooling, model compression tricks and how to think practically about AI integration. If you want to ship something smart without selling your soul to the big AI model providers, this talk is for you.
From AI Hype to Production Reality: Lessons for Developers
Many AI projects fail not because of weak models but due to overlooked fundamentals in design, evaluation, and deployment. Using real-world cases, this talk walks through the pitfalls of rushing from PoC to production: high costs, hallucinations, security risks, and lack of evaluation. Developers will learn practical strategies including prompt engineering, observability, guardrails, and cost control to build production-ready AI systems.
Building AI That Learns and Adapts: A Case Study in MRI Diagnostics
Explore how AI and continual learning can revolutionize MRI diagnostics, using our real-world case study in detecting Focal Cortical Dysplasias (FCD)—a crucial factor in epilepsy treatment. In this session, we’ll dive into how continual learning techniques, inspired by human adaptability, help AI models improve diagnostic accuracy, minimize false positives, and handle evolving data in medical settings. We’ll discuss challenges like catastrophic forgetting, maintaining model performance in dynamic environments, and practical strategies that developers can apply to other domains. Gain hands-on insights into building resilient AI systems that evolve and adapt to new data, ensuring long-term reliability in critical applications.
NDC Oslo 2025 Sessionize Event
NDC Melbourne 2025 Sessionize Event
Swetugg Stockholm 2025 Sessionize Event
Codemotion Milan 2024 Sessionize Event
DevConf 2024 Sessionize Event
Data Science Summit Machine Learning Edition 2024
Special edition of the most important Data Science event in Poland dedicated to Machine Learning
Session Title: Focal Cortical Displasias and how to find them (on MRIs))
Webinale Berlin 2024
The holistic web conference
Webinale is the conference for product owners, web designers, and frontend coders. It combines user experience with code and offers a variety of topics that are important for digital professionals nowadays.
Session: Lessons Learned from Implementing Generative AI Solutions
DevConf 2023 Sessionize Event
Data Science Summit Machine Learning Edition 2023
Special edition of the most important Data Science event in Poland dedicated to Machine Learning
Session Title: From Sandbox to Real World: Challenges of Implementing Generative AI Solutions
Re-Work: AI Assistant Summit London 2019
Session title: Answering E-Mails in No Time – Short Story of How to Use NLP in Modern Enterprises
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