Dementia Diagnostics Platform
Nullwest audited and rebuilt BrainDoc's medical diagnostics platform — closing security holes, raising AI transcription accuracy from 60% to 90%+, and fixing 74 bugs — turning a fragile prototype into a demo-ready product.
Software in the middle of a clinical decision
BrainDoc, founded by Dr. Mitch and Dr. Emily Clionsky, is a computerized diagnostics and report-generation platform. A patient takes a guided, video-driven cognitive assessment; the platform transcribes and scores their spoken responses; the clinician gets a report they can act on.
A scoring error isn't a bug ticket — it's a wrong number in front of a doctor.
X-ray first. Then fix everything.
BrainDoc arrived with an inherited five-layer Next.js app and a sense it wasn't ready for what came next. Before changing a line, we ran an x-ray audit of the full system — then fixed it, end to end.
Security overhaul
Authentication on every sensitive endpoint, cross-user tampering fixed, PII scrubbed from logs, prompt-injection defenses and rate limiting added.
AI scoring rebuilt
Rewrote scoring prompts, added Zod schema validation in CI, and validated recognition against 500 AI-generated voice responses.
Smarter model economics
Re-routed scoring through OpenRouter to a smaller, better-performing model — cutting per-test AI cost without sacrificing quality.
Infrastructure migration
Moved hosting from Vercel to AWS for stability, lower cost, and room to scale.
Reporting consolidated
Replaced 30 separate PDF reports with one HTML-generated report — thirty maintenance surfaces collapsed into one.
Product polish
74 bugs fixed, video prefetch and preload for performance, and code-quality hardening throughout.
Accuracy you can measure
We didn't swap models and hope. Rewritten prompts, Zod validation on every AI response, and a headless-Chrome harness testing 500 synthetic voices mean accuracy stays measurable — not anecdotal.
Medical software does not get to leak
Patient reports sat in unauthenticated PDFs and PII landed in logs. We locked every sensitive endpoint and treated a health platform like the regulated-adjacent product it is — before anyone forced the issue.
Inherited code is a liability until proven otherwise
The x-ray audit came first, so every decision — AWS migration, model re-routing, report consolidation — traced to a documented business reason. The founders left understanding their own platform.
Every SOW item delivered — and then some
Work shipped at no extra cost included items explicitly excluded from the SOW: the full PDF-to-HTML report redesign, sectioned scoring, UI cleanup across the site, and the 500-response voice test harness. When late nav bugs surfaced outside Phase 2 scope, we fixed those free too.
BrainDoc now has a platform its founders can confidently demo to clinicians and investors alike.
Beyond the SOW, at no charge
Inherited a codebase you're not sure you can trust?
We've been auditing, securing, and shipping production AI systems from Apple to Xembly to platforms like BrainDoc. Let's talk about yours.