London / Online • Journal-club style

Applied Deep Learning

A discussion-led community for practitioners who want to understand how widely used deep learning models actually work: their assumptions, optimisation choices, inductive biases, and real-world failure modes.

Join / RSVP How sessions run
Focus: established papers and “standard” models people use every day, explained clearly and debated constructively.
Applied Deep Learning logo
Tip: Meetup crops profile images—prefer a square logo (e.g., 1024×1024).

About the group

This meetup is for engineers, data scientists, researchers, and technically curious builders applying deep learning in practice. The aim is to move beyond “how to use a model” and towards understanding why it works, where it fails, and what it is implicitly assuming about data and the world. This is an excellent space for deep learning practioners and interested parties to grow their skills and to develop their community.

What this group is (and isn’t)
  • Is: a structured, discussion-led journal club for established, high-impact models and methods.
  • Isn’t: a hype-driven “latest arXiv” feed, a sales channel, or a beginner-only tutorial series.

Basic framework

Sessions are typically 60 minutes online (Zoom or Google Meet), designed to be practical, repeatable, and time-respectful.

Part 1 — Curated introduction (≈ 30 min)
  • Problem framing: what the method/model was trying to solve
  • Core figures and design choices (architecture, loss, training setup)
  • Key assumptions and inductive biases
  • Where it tends to work well, and why
Part 2 — Discussion & critique (≈ 30 min)
  • Open critique of the framing and interpretation
  • Failure modes and “gotchas” from real deployments
  • Alternative viewpoints and extensions
  • Practical takeaways: what to remember when using it
Sessions emphasise intuition and visual understanding over heavy derivations, while welcoming technical depth and diverse perspectives.

Example topic areas

Topics vary by interest and facilitator, but we generally focus on established methods that are widely used in the field.

Vision & detection/segmentation Signal processing & time series Inductive biases & representations Optimisation, losses, calibration Generative models & diffusion Physics-/simulation-driven ML Real-time inference & deployment
Want to present?
Members are welcome to facilitate sessions. A good session is not a perfect lecture—it's a well-structured reading plus honest discussion of assumptions and failure modes.

Join / RSVP

The MeetUp page:

Meetup page Events Listing Community chat
Expect a practical, respectful environment. No hype, no sales, and no pressure to speak—though questions and critique are encouraged.

Organiser

Dr. Dominic Waithe is a lead engineer and data scientist with a background spanning biophysics, signal processing, computer vision, and applied deep learning. Since 2021 he has worked in industry developing and deploying machine learning systems for physics- and imaging-driven problems, including real-time models on edge hardware and cloud-based platforms. Alongside his work in machine learning, Dominic has a long-standing interest in signal processing and visualisation, which culminated in the development of Sound to Vision (soundtovision.com), a browser-based platform for creating audio-reactive visuals for live performance, streaming, and video. His freelance and creative-technical work is showcased at odlogo.co.uk.

odlogo.co.uk soundtovision.com GitHub YouTube LinkedIn X
Please reach out to Dominic using the meetUp direct messaging or through social media direct messaging.