Company profile · 3 min read
Aluna: curated oncology data for biomedical AI
Aluna builds curated oncology datasets and evaluation suites across pathology, genomics and clinical data for health-AI teams.
Cofounder of Lazyweb (a16z Speedrun). Previously grew subscriptions and built native ads at Duolingo. Writes First1000 (90,000+ subscribers).
Published · Updated
What it does
Aluna builds curated biomedical datasets and evaluation suites for health AI, currently focused on oncology. Its current homepage exposes OncoBench, a planned oncology benchmarking suite, PathFoundry, which the company describes as 2M+ whole-slide images across 25+ tumor types with pathologist annotations, and Duplex, a multimodal dataset pairing whole-genome sequencing, clinical information and digitized pathology slides.
The fit is an AI lab, biotech team or clinical-model builder that needs more than a public benchmark or a raw data dump. Aluna’s value depends on curation, provenance, annotation quality and the ability to compare models across cancer types and treatment modalities. The buyer should be prepared to ask how the data was collected and what the evaluation actually measures.
Why I’d look closer
The product focus is narrow enough to be useful. Oncology datasets can be structured around a real clinical question rather than a generic “healthcare AI” label, and the homepage’s combination of imaging, genomics and clinical data points toward multimodal evaluation. The page says PathFoundry includes H&E, IHC and IF stains with diagnoses and annotations from board-certified pathologists; that is a company description, not an independent quality audit.
The official YC profile describes David Chu as a Brown computer-science/economics graduate working with health data. The same profile carries an older Company Launches entry for Fortress, a separate database-orchestration product. I am treating the current Aluna directory description and matching biomedical homepage as the current product signal, and not carrying Fortress features into this profile.
What I’d ask
What is available today versus “coming soon,” and can a buyer inspect data dictionaries, inclusion criteria, provenance, licensing and annotation agreements before committing? I’d request an oncology benchmark sample, a held-out evaluation protocol and the error taxonomy for pathology and clinical labels. The sources checked did not expose public pricing, a docs surface or an independent review, so the first engagement should be a data-quality and rights diligence exercise.
My editorial take
Aluna is a good fit for teams that need oncology-specific evidence and can tolerate a curated, consultation-led data relationship. The current website gives enough specificity to warrant a conversation. The practical decision is whether the data is usable for the buyer’s exact model and regulatory context, not whether the dataset headline is large.
Quick facts
| Field | Sourced detail |
|---|---|
| Buyer fit | Biomedical AI labs, biotech and clinical-model teams |
| Current focus | Oncology datasets and benchmarking |
| Data examples | Pathology, whole-genome sequencing and clinical information; company-described |
| Public pricing | Not exposed in the sources checked |
Sources checked
Checked 2026-09-19.
| Source | Used for |
|---|---|
| YC company profile | Current product description, founder and older launch-context distinction |
| Aluna homepage | Current dataset names, oncology scope and company-described data attributes |
Cohort context
Aluna is listed in Summer 2024. In our 2026-09-18 directory snapshot, 161 of 248 listed companies in that cohort have YC’s primary industry label B2B (64.9%). This is a current-directory comparison, not an original intake count or a performance ranking. Nine-cohort dataset.
Public website snapshot
Observed 2026-09-19T16:16:47.495Z in raw homepage HTML. This records visible metadata and advertised links, not agent execution or product quality.
| Signal | Homepage observation |
|---|---|
| Product description metadata | Observed |
| Canonical link | Not observed in this response |
| H1 or H2 heading | Observed |
| Typed structured data | Not observed in this response |
| Docs/developer link | Not observed in this response |
| Pricing link | Not observed in this response |
| llms.txt link | Not observed in this response |
| Markdown alternate | Not observed in this response |
Public observations · Collection method. Missing links here do not establish that a capability or file is absent elsewhere.
About the author
Cofounder of Lazyweb (a16z Speedrun). Previously grew subscriptions and built native ads at Duolingo. Writes First1000 (90,000+ subscribers).
I cofound Lazyweb and publish Mudpie. This is an owner-written publication, not an independent testing organization. Research notes distinguish observations, sourced reporting and editorial judgment.
First1000 ↗ · X ↗