AI Engineering Podcast From MRI to World Models: How AI Is Changing What We See
Daniel Sodickson, Chief of Innovation in Radiology at NYU Grossman School of Medicine, shares his expertise in AI and medical imaging. He unveils the evolution from linear MRI to deep learning, emphasizing the distinction between upstream AI that influences measurement and downstream AI that interprets images. Their discussion includes the challenges of cross-disciplinary knowledge, ethical implications of decoding brain activity, and innovative concepts like 'imaging without images.' Daniel highlights the necessity of human oversight as AI transforms healthcare and visual understanding.
48:51
Early MRI Deep Learning Failure And Fix
- Daniel Sodickson described early deep learning MRI outputs that looked sharp but were "completely unacceptable" to radiologists.
- Iterating with radiologists and physics knowledge made the images indistinguishable from traditional scans.
MRI And Radio Astronomy Share Equations
- Sodickson found identical math underlying MRI and radio astronomy image formation.
- He describes it as discovering a long-lost sibling despite different notations.
Upstream AI Rewrites What We Measure
- Sodickson divides imaging AI into downstream (interpretation) and upstream (changing measurement).
- Upstream AI can redesign sensing to make imaging faster, cheaper, and more accessible.
Get the Snipd Podcast app to discover more snips from this episode
Get the app 1 chevron_right 2 chevron_right 3 chevron_right 4 chevron_right 5 chevron_right 6 chevron_right 7 chevron_right 8 chevron_right 9 chevron_right 10 chevron_right 11 chevron_right 12 chevron_right 13 chevron_right 14 chevron_right 15 chevron_right 16 chevron_right 17 chevron_right 18 chevron_right 19 chevron_right 20 chevron_right 21 chevron_right 22 chevron_right 23 chevron_right 24 chevron_right 25 chevron_right 26 chevron_right
Intro
00:00 • 3min
Dan introduces his background in imaging physics
02:30 • 26sec
Evolution from linear reconstruction to deep learning
02:57 • 1min
What is an image and its role in AI?
03:59 • 2min
Challenges of different imaging modalities for AI
05:45 • 1min
Cross-domain similarities between imaging fields
07:12 • 4min
AI as a translator across scientific disciplines
10:50 • 50sec
Downstream versus upstream AI in imaging
11:40 • 2min
Upstream AI enabling faster, cheaper imaging
13:49 • 2min
Human judgment remains central alongside AI
15:59 • 3min
World models over task-specific vision systems
18:39 • 2min
Multimodal representation learning beyond LLMs
20:59 • 3min
Cross-modality masked autoencoding for richer semantics
23:46 • 1min
Explainability strategies for downstream and upstream AI
25:10 • 3min
AI-driven continuous health monitoring use case
27:41 • 2min
Making continuous medical monitoring affordable and scalable
29:36 • 3min
AI agents as triage for clinicians
32:56 • 36sec
Remote sensing and astronomy benefit from always-on agents
33:32 • 2min
Decoding brain activity and ethical implications
35:21 • 3min
Designing imaging for machines instead of humans
38:14 • 3min
Information lost when translating raw signals to human images
40:53 • 17sec
Lessons learned: biology-inspired architectures and world models
41:10 • 1min
Where not to rely solely on plug-and-play models
42:32 • 2min
Preserving truth as our vision changes
44:06 • 2min
Gaps in tooling: distilling core ML methodologies
46:20 • 2min
Outro
47:50 • 55sec
Summary
In this episode of the AI Engineering Podcast Daniel Sodickson, Chief of Innovation in Radiology at NYU Grossman School of Medicine, talks about harnessing AI systems to truly understand images and revolutionize science and healthcare. Dan shares his journey from linear reconstruction to early deep learning for accelerated MRI, highlighting the importance of domain expertise when adapting models to specialized modalities. He explores "upstream" AI that changes what and how we measure, using physics-guided networks, prior knowledge, and personal baselines to enable faster, cheaper, and more accessible imaging. The conversation covers multimodal world models, cross-disciplinary translation, explainability, and a future where agents flag abnormalities while humans apply judgment, as well as provocative frontiers like "imaging without images," continuous health monitoring, and decoding brain activity. Dan stresses the need to preserve truth, context, and human oversight in AI-driven imaging, and calls for tools that distill core methodologies across disciplines to accelerate understanding and progress.
Announcements
Parting Question
In this episode of the AI Engineering Podcast Daniel Sodickson, Chief of Innovation in Radiology at NYU Grossman School of Medicine, talks about harnessing AI systems to truly understand images and revolutionize science and healthcare. Dan shares his journey from linear reconstruction to early deep learning for accelerated MRI, highlighting the importance of domain expertise when adapting models to specialized modalities. He explores "upstream" AI that changes what and how we measure, using physics-guided networks, prior knowledge, and personal baselines to enable faster, cheaper, and more accessible imaging. The conversation covers multimodal world models, cross-disciplinary translation, explainability, and a future where agents flag abnormalities while humans apply judgment, as well as provocative frontiers like "imaging without images," continuous health monitoring, and decoding brain activity. Dan stresses the need to preserve truth, context, and human oversight in AI-driven imaging, and calls for tools that distill core methodologies across disciplines to accelerate understanding and progress.
Announcements
- Hello and welcome to the AI Engineering Podcast, your guide to the fast-moving world of building scalable and maintainable AI systems
- When ML teams try to run complex workflows through traditional orchestration tools, they hit walls. Cash App discovered this with their fraud detection models - they needed flexible compute, isolated environments, and seamless data exchange between workflows, but their existing tools couldn't deliver. That's why Cash App rely on Prefect. Now their ML workflows run on whatever infrastructure each model needs across Google Cloud, AWS, and Databricks. Custom packages stay isolated. Model outputs flow seamlessly between workflows. Companies like Whoop and 1Password also trust Prefect for their critical workflows. But Prefect didn't stop there. They just launched FastMCP - production-ready infrastructure for AI tools. You get Prefect's orchestration plus instant OAuth, serverless scaling, and blazing-fast Python execution. Deploy your AI tools once, connect to Claude, Cursor, or any MCP client. No more building auth flows or managing servers. Prefect orchestrates your ML pipeline. FastMCP handles your AI tool infrastructure. See what Prefect and Fast MCP can do for your AI workflows at aiengineeringpodcast.com/prefect today.
- Unlock the full potential of your AI workloads with a seamless and composable data infrastructure. Bruin is an open source framework that streamlines integration from the command line, allowing you to focus on what matters most - building intelligent systems. Write Python code for your business logic, and let Bruin handle the heavy lifting of data movement, lineage tracking, data quality monitoring, and governance enforcement. With native support for ML/AI workloads, Bruin empowers data teams to deliver faster, more reliable, and scalable AI solutions. Harness Bruin's connectors for hundreds of platforms, including popular machine learning frameworks like TensorFlow and PyTorch. Build end-to-end AI workflows that integrate seamlessly with your existing tech stack. Join the ranks of forward-thinking organizations that are revolutionizing their data engineering with Bruin. Get started today at aiengineeringpodcast.com/bruin, and for dbt Cloud customers, enjoy a $1,000 credit to migrate to Bruin Cloud.
- Your host is Tobias Macey and today I'm interviewing Daniel Sodickson about the impact and applications of AI that is capable of image understanding
- Introduction
- How did you get involved in machine learning?
- Images and vision are concepts that we understand intuitively, but which have a large potential semantic range. How would you characterize the scope and application of imagery in the context of AI and other autonomous technologies?
- Can you give an overview of the current state of image/vision capabilities in AI systems?
- A predominant application of machine vision has been for object recognition/tracking. How are advances in AI changing the range of problems that can be solved with computer vision systems?
- A substantial amount of work has been done on processing of images such as the digital pictures taken by smartphones. As you move to other types of image data, particularly in non-visible light ranges, what are the areas of similarity and in what ways do we need to develop new processing/analysis techniques?
- What are some of the ways that AI systems will change the ways that we conceive of
- What are the most interesting, innovative, or unexpected ways that you have seen AI vision used?
- What are the most interesting, unexpected, or challenging lessons that you have learned while working on imaging technologies and techniques?
- When is AI the wrong choice for vision/imaging applications?
- What are your predictions for the future of AI image understanding?
Parting Question
- From your perspective, what are the biggest gaps in tooling, technology, or training for AI systems today?
- Thank you for listening! Don't forget to check out our other shows. The Data Engineering Podcast covers the latest on modern data management. Podcast.__init__ covers the Python language, its community, and the innovative ways it is being used.
- Visit the site to subscribe to the show, sign up for the mailing list, and read the show notes.
- If you've learned something or tried out a project from the show then tell us about it! Email hosts@aiengineeringpodcast.com with your story.
- To help other people find the show please leave a review on iTunes and tell your friends and co-workers.
- MRI == Magnetic Resonance Imaging
- Linear Algorithm
- Non-Linear Algorithm
- Compressed Sensing
- Dictionary Learning Algorithm
- Deep Learning
- CT Scan
- Cambrian Explosion
- LIDAR Point Cloud
- Synthetic Aperture Radar
- Geoffrey Hinton
- Co-Intelligence by Ethan Mollick (affiliate link)
- Tomography
- X-Ray Crystallography
- CERN
- CLIP Model
- Physics-Guided Neural Network
- Functional MRI
- A Path Toward Autonomous Machine Intelligence by Yann LeCun