Practical Functional AI
Applied AI systems built from real-world problems.
AI Thought Lab is the applied AI workshop of Richard Coulson — a place for building, testing, and documenting practical AI systems across computer vision, automation, retrieval, and physical-world workflows.
Current Focus
- Image-based fabric identification
- OCR and label recognition
- Vector search and retrieval
- Factory and warehouse workflow tools
Featured Projects
Real systems, active prototypes, and practical AI workflows.
The center of AI Thought Lab is proof-of-work: projects that connect AI techniques to physical-world problems, messy inputs, and practical user workflows.
Case Study
AI-powered Fabric Identification
Swatch ID
AI-powered fabric identification using visual embeddings and local catalog retrieval.
View Case StudyWorking Prototype
Computer Vision / OCR
Bolt Finder
An OCR-first computer vision prototype for scanning warehouse-style rack images, extracting roll-number candidates, and highlighting matching fabric roll labels.
View Case StudyCase Study
Manufacturing Intelligence
Cutting Room Explorer
Evidence-led manufacturing intelligence for exploring cutting-room activity, investigating machine and marker workload, and tracing findings to individual production records.
View Case StudyWhat This Lab Builds
AI tools that survive contact with real workflows.
The focus is not novelty AI. It is practical systems work: computer vision, retrieval, automation, OCR, local-first tools, and human-in-the-loop workflows built around real constraints.
Computer Vision
Image-based identification, visual matching, OCR, and label recognition.
Retrieval Systems
Vector search, embeddings, catalog lookup, and RAG-style workflows.
Workflow Tools
Human-centered AI utilities for factories, warehouses, samples, and process work.
Local Practical AI
Systems designed around useful results, privacy, speed, and real deployment constraints.
Proof Over Hype
Built from observation, tested against constraints, refined through use.
AI Thought Lab exists to document useful AI systems as they are built: what works, what breaks, what has to be simplified, and what becomes valuable once the tool meets the real world.
Lab Notes
Tech observations.
Human Review Is a Feature, Not a Failure.
Confidence, explainability, and traceability are most useful when AI supports informed human judgment rather than replacing it.
EngineeringProducts Tell Stories. Projects Tell Histories.
Why product pages should explain present value while engineering records preserve the path that produced it.
Build Log
Current progress.
AI Thought Lab Projects Become a Software Showcase
Redesigned the Projects page as a visual portfolio that distinguishes polished case studies from working prototypes.
Swatch IDSwatch ID Gains a Focused Product Demonstration
Reorganized Swatch ID around a concise video demonstration and a human-reviewed visual-search workflow for upholstery fabrics.
Built by Richard Coulson
Applied AI from the perspective of a practical systems builder.
Richard brings together R&D experience, CAD and pattern workflow knowledge, factory-floor observation, automation thinking, and hands-on AI prototyping. AI Thought Lab is the public record of that work.