2025-10-30 –, Auditorium
What if you could look into a magical mirror and see your deepest desires, generated by Python and AI? In this talk, we’ll build Harry Potter's Mirror of Erised using LLMs and text-to-image models to visualize personal dreams.
What if the Mirror of Erised from Harry Potter actually existed and you could look into it and see your deepest desires visualized? What if you could build it using Python?
In this talk, I’ll introduce a creative, AI-powered project that transforms a user's personal hopes and dreams into a visual representation using NLP and image generation. The user writes freely about what they long for. A large language model (LLM) analyzes the text to extract emotional tone, symbols, and core themes. Then, this interpretation is converted into a visual prompt for a text-to-image model, which generates an image.
The final image is dynamically displayed behind the user's webcam feed, turning their screen into a personalized, digital “Mirror of Erised.”
I’ll walk through the full technical pipeline: from prompt engineering and emotion detection to visual generation and building the interactive UI with Python. This talk combines storytelling with code and aims to inspire developers to explore how Python and AI can bring deeply personal, emotional, and magical experiences to life.
I'm Özge, a curious soul with a computer engineering degree and a heart full of ideas. I'm currently shaping the future as an AI Research Engineer at Huawei. I work in AI research, but I’m just as passionate about blending creativity with code. Whether it’s turning emotions into visuals, building fictional chatbots, or crafting data stories, I love making tech feel personal. I write, build, explore, and sometimes get beautifully lost in too many ideas, but always with Python by my side.