Workshop w/ Factored-Beyond Matrix Factorization: Deep RecSys Architectures in Action
- Date
- 2025-08-22
- Host
- Colombia Tech Week
About this event
If you care about how recommendation systems actually work beyond the standard matrix factorization playbook, this workshop is built for you. Workshop w/ Factored-Beyond Matrix Factorization: Deep RecSys Architectures in Action is a hands-on, in-person session focused on what comes next when classic collaborative filtering stops being enough. This is the kind of event where technical curiosity meets practical discussion. You’ll spend time with people who want to understand modern recommender system design more deeply, compare approaches, and talk through how deep architectures change the way we think about ranking, relevance, and user-item interaction. About the Event This workshop centers on a clear theme: moving beyond traditional matrix factorization into deeper recommendation system architectures that can capture richer patterns in behavior, context, and content. Rather than treating recommender systems as a black box, the session is designed to make the underlying ideas more legible and actionable. Expect a format that feels more interactive than a passive talk. The workshop framing suggests a session where attendees can engage with concepts, discuss tradeoffs, and connect ideas to real system design questions instead of only watching slides from a distance. Because the topic sits at the intersection of machine learning, product thinking, and applied engineering, the room is likely to attract a mix of practitioners and curious learners. That mix is a strength: it creates space for both technical depth and grounded conversation about where these methods matter in practice. Whether you come from recommender systems specifically or from adjacent ML work, the event offers a focused opportunity to think carefully about architecture choices, model behavior, and the practical meaning of “deep” in modern RecSys work. What to Expect You should expect a session anchored in deep recommendation system concepts, with emphasis on architectures that extend or outperform baseline matrix factorization approaches in more complex settings. The title points toward applied discussion, so the value here is not just theory, but seeing how these ideas show up in real modeling decisions. Likely areas of focus include: How matrix factorization works well—and where it starts to break down What deep recommendation architectures add in terms of representation learning and nonlinear interaction modeling How to think about user and item signals when behavior is sparse, messy, or multi-dimensional Architectural tradeoffs between simplicity, expressiveness, interpretability, and operational complexity Practical discussion with other attendees interested in recommender systems, ML systems, and applied modeling Because this is an in-person workshop, you can also expect more direct interaction than you’d get from a webinar or recorded session. That may mean asking questions in real time, comparing notes with peers, or digging into the reasoning behind different modeling choices. There is also a strong community angle here. With tags spanning community, networking, meetup, and social, this is not only a technical session; it is also a place to meet people who are actively thinking about recommendation systems and adjacent machine learning problems. Why Attend If you’ve mostly encountered recommendation systems through matrix factorization, this workshop gives you a sharper view of the broader design space. It helps bridge the gap between foundational methods and the deeper architectures increasingly used when teams need more flexible, expressive models. You’ll leave with a better sense of when classic approaches are enough, when they are not, and what alternatives are worth understanding. That kind of judgment is useful whether you are building a recommender, evaluating models, or simply trying to follow modern RecSys literature with more confidence. This event is also valuable because it combines learning with conversation. Workshops create room for the kinds of questions that matter in real work: What problem does a deeper architecture actually solve? What new complexity does it introduce? How do you balance performance gains against implementation and maintenance costs? What should you pay attention to when choosing among RecSys approaches? Beyond the technical content, there is clear value in being in the room with others who care about applied ML. If you want to expand your network in a way that feels relevant to your interests, this event offers a focused context for doing that. Practical Details This is an in-person event, which makes it a good fit for attendees who prefer live discussion, stronger attention, and easier networking before or after the session. If you learn best by being physically present and able to interact directly, that format is part of the appeal. The workshop takes place on Friday, August 22 at 1:00 PM GMT-5. A midday Friday slot works well if you want to carve out dedicated time for learning and conversation without turning it into an all-day commitment. A few useful planning notes: Format: Workshop-style, with a technical focus and community feel Location type: In person Theme: Deep recommender system architectures beyond matrix factorization Audience energy: Likely a mix of learning, discussion, and networking If this topic sits anywhere near your work or curiosity, it is worth making time for. You’ll get a more current view of recommender system thinking, a better vocabulary for discussing model architecture, and a room full of people who are interested in the same questions.
Who should attend
This will feel especially worthwhile if you want a more practical, current understanding of recommender systems and enjoy learning in a room with other technically curious people. - You work in **machine learning, data science, or software engineering** and want to better understand how modern recommendation approaches go beyond classic matrix factorization. - You’re building, evaluating, or supporting **ranking, personalization, or recommendation features** and want stronger intuition about architectural choices. - You’ve studied the basics of collaborative filtering and now want to understand **what deep RecSys architectures actually change** in practice. - You like events that combine **technical learning with real conversation**, rather than purely passive listening. - You’re interested in **applied ML communities** and want to meet people who care about recommender systems, modeling tradeoffs, and production-minded thinking. - You’re simply RecSys-curious and want a structured, approachable way to get closer to the topic without needing a highly formal conference setting.