Model Compression & Optimization: Making AI Models Faster | #GirlsWhoML

How do you take a state-of-the-art AI model and make it small enough, fast enough, and cheap enough to actually ship? In this Workshop Wednesdays session — part of the GirlsWhoML × Mentor Me Collective series — Staff AI Research Scientist Dominika Woszczyk (Iconic) breaks down the techniques that turn research-grade models into production-ready systems. We'll explore the core methods practitioners use to compress and optimize modern AI models, including: → Quantization — reducing precision without losing performance → Pruning — removing what your model doesn't need → Knowledge distillation — teaching smaller models to behave like larger ones → Practical trade-offs between latency, accuracy, memory, and cost → Real-world considerations for deploying optimized models at scale Whether you're an ML engineer trying to cut inference costs, a researcher curious about efficient AI, or a student preparing to work on production systems, you'll leave with a clearer mental model for how and when to compress. About the speaker Dominika Woszczyk is a Staff AI Research Scientist at Iconic, with a PhD from Imperial College London where her work spanned speech AI, privacy-preserving ML, and knowledge distillation from large language models. Her research has been published at venues including PoPETs and ACSAC. About the series Workshop Wednesdays is a free, live learning series. This session is co-produced by GirlsWhoML and Mentor Me Collective, bringing technical women practitioners together to teach the skills shaping modern AI. New sessions every Wednesday. 🗓 Originally streamed Wednesday, May 27, 2026 🕑 2 PM EST / 7 PM UK / 12:30 AM IST Connect with Mentor Me Collective 🌐 mentormecollective.org 💼 LinkedIn: Mentor Me Collective 📩 Subscribe for upcoming Workshop Wednesdays sessions #MachineLearning #AI #ModelCompression #Quantization #KnowledgeDistillation #MLEngineering #WomenInTech #GirlsWhoML #MentorMeCollective #WorkshopWednesdays