I’m excited to share that I was recently part of a research paper published on ScienceDirect. This work explores how interactive, browser-based data visualisation tools can improve the way clinicians and researchers engage with complex clinical trial datasets. In particular, we introduce VisionOnc, a platform designed to allow users to dynamically explore patient-level oncology data—moving … [Read more…]
Introduction Slurm (Simple Linux Utility for Resource Management) is an open-source workload manager and job scheduler designed for Linux-based high-performance computing environments. Its primary role is to allocate compute resources, queue and schedule jobs, and enforce fair-use policies across shared infrastructure such as CPU clusters, GPU nodes, and large-memory systems. Rather than relying on ad-hoc … [Read more…]
The Orchestrator: Owning the Loop In the previous article, we established a simple but often misunderstood idea: an agent is not an autonomous entity. An agent is a loop. It observes input, reasons about the current situation, proposes a next step, executes that step through predefined mechanisms, observes the result, and then either continues or … [Read more…]
Agentic AI has quickly become one of those terms that carries more weight than clarity. Depending on where you encounter it, it is either framed as the next inevitable step toward autonomous systems or dismissed as unnecessary complexity layered on top of language models. Both views miss the point. Agentic AI is neither magical nor … [Read more…]
Introduction In a previous article, we explored a zero-code approach to storing DICOM metadata by persisting vendor-specific and modality-specific tags as JSONB in an extradicomtags column. This approach embraces the reality of clinical imaging data: the schema is heterogeneous, nested, and constantly evolving. By storing the raw metadata as a JSON document and applying GIN … [Read more…]
The Real Problem: Ingestion Is Not the Product In medical imaging platforms, ingestion is often treated as a prerequisite rather than a core problem. Yet teams repeatedly invest months building custom pipelines just to move DICOM files from point A to point B. A typical from-scratch ingestion pipeline ends up requiring filesystem watchers or DICOM … [Read more…]
Clinical environments generate vast amounts of operational data from electronic medical records, oncology information systems, PACS archives, and imaging and treatment-planning platforms. This data has enormous value for research, AI model development, and quality improvement. Yet pulling this data from production systems in a form suitable for analysis is exceptionally difficult. The core challenge is … [Read more…]