Learning and documenting pathology informatics Designing responsible histopathology workflows
I am Dr. Fernando Soto, a physician and computer science graduate preparing for anatomical pathology residency. Histopath.ai is my public notebook: a living study path, annotated case library, and running project log for the pathology informatics work I build along the way. My current focus is whole-slide imaging pipelines, AI validation, and report-style pathology writing.
- Training
- MD and Computer Science graduate
- Focus
- Pathology informatics, WSI pipelines, AI validation and QA
- Current projects
- Breast carcinoma WSI pipeline, DCP 601–602 notes, ML metrics notebook
- Goal
- Anatomical pathology residency with emphasis on pathology informatics
About
Who I am, why I built histopath.ai, and how to reach me.
Visit sectionCase Library
Educational case reports emphasising terminology, diagnostic reasoning, and QA follow-ups.
Visit sectionReference guide
Quick-reference docs for datasets, tissue guides, and mathematics refreshers.
Visit sectionProjects
Deep learning, automation, and data projects that bridge histology practice with engineering.
Visit sectionReference docs
Latest reference documents
Freshly updated articles from docs.histopath.ai. New pages will appear here automatically as they’re added to the reference guide.
11/25/2025
Calculus
Derivatives, chain rule, and loss surfaces for optimisation in models.
11/25/2025
Linear Algebra
Vectors, matrices, decompositions—applied to stain normalisation and embeddings.
11/25/2025
Probability
Distributions, expectation/variance, and sampling plans for QC and reader studies.
11/25/2025
Statistics
Hypothesis testing, confidence intervals, and evaluation curves for patch/slide models.
11/25/2025
Step 6 – ML & Modeling (The Diagnosis Phase)
Classical ML with scikit-learn on tabular features and deep learning with PyTorch on raw images, side by side.
11/25/2025
ML & Modeling
Use patches, features, and annotations to train models for classification, prediction, and triage.
Latest cases
Practice cases with report-style write-ups
Every case captures diagnostic language, QA follow-ups, and the questions I bring to attending pathologists.
Featured project
Breast Carcinoma WSI Pipeline
Reproducible workflow for tiling, training, and reporting attention-based slide classifiers on invasive breast carcinoma.
- Standardised patch extraction and stain normalisation using reproducible configs.
- Slide-level attention pooling with integrated attribution overlays for case review.
- Validation dashboards aligned with CAP/CLIA quality metrics and internal QA checks.
Notebook updates
What I’m reading and shipping
Short reflections on papers, conferences, and project milestones that influence the study path. Updated as I learn, usually monthly.
View notebook archive →Let’s compare notes
I welcome feedback, mentorship, and collaboration opportunities, especially from teams building pathology informatics infrastructure or evaluating AI tools. Reach out if you spot gaps, want to share material, or a partner on validation work.
- fsoto@histopath.ai
- GitHub
- github.com/drsoma
- ORCID
- 0009-0007-5857-3476