Notes
Writing on machine learning, medical AI, and research.
- LLM-Generated Code Should Be Judged by Its Blast Radius
- When Agents Start Talking
- Chain-of-Thought Is Becoming a Crutch
- Humankind in the Age of AI
- AI Job Displacement: My take by analysing 30 Datasets
- 350 Devices in One Year: Review of 2025 FDA AI/ML List
- LLM Surveillance and Anthropic’s Redline
- The Evolution of Search: From Exact Matches to True Understanding
- Weather Forecasting & Broken Medical VLMs
- The Digital Eraser: Why Teaching Models to Forget is as Important as Learning
- The Invisible Hand in the Age of AI
- A highly Opinionated View: How LLMs Will Rewire Product Development in MedTech-Manufacturing
- Circuit Tracing: Finding Medical Features in Gemma 3
- Your LLM Medical Device Just Got Smarter. Does the FDA Need to Know?
- What Does Medical VLM Actually See? Experiments with MedGemma and Sparse Autoencoders
- Opening the Black Box: How to See What Your Vision Language Model is Actually Looking At
- Fluent But Wrong: LLM and Healthcare
- When AI Radiologists Get Confused: The Critical Challenge of VLM Robustness in Medical Diagnostics
- Building AI Agents with Multimodal Models: The Final Challenge
- Building AI Agents with Multimodal Models: Part 4
- Building AI Agents with Multimodal Models: Part 3
- Building AI Agents with Multimodal Models: Part 2
- BitNet b1.58: The Death of Multiplication
- Building AI Agents with Multimodal Models : Part 1
- OCR on Engineering Drawings with a 0.9B Vision-Language Model
- Why a 0.9B VLM can be a serious OCR engine
- Multimodal LLMs in Healthcare: What’s Actually Working
- A guide to LLM evaluation metrics
- Data Generating Process
- Understanding Random Variables: A Practical Guide for Engineers
- Bayesian Optimization
- Machine Learning Primer