Kernel Health
Amara Okafor
@amara_builds
On-device diagnostic models for low-connectivity clinics.
Health AILagos, NGPre-seedOUTBOUNDCOLD START
Legibility-penalized genius — weak English, elite substance
Founder Score
A running score for this person, not this pitch — it persists across applications and never resets to zero.
40Feb 10, 25
Surfaced via hackathon win + on-device ML repo
61Apr 22, 25
Legibility-Bias Corrector applied → substance revealed
73Jun 15, 25
Deployed pilot in 2 clinics (verified)
3-Axis Score
Founder, market, and idea-fit are scored independently and never averaged — disagreement between them is signal, not noise.
Founder
Is this person capable of pulling this off?
76improving
Substance is elite (novel quantization approach shipped, real clinic pilot). Naive LLM scoring under-rated her by 17 pts purely on English fluency; the Legibility-Bias Corrector removed that penalty.
2 evidence points backing this score
Market
Is this a market worth building a fund-scale outcome in?
64improving
Low-connectivity clinical AI is under-served; grant + payer tailwinds. SOM modest near-term, large by 2030.
Idea vs Market
Does the specific idea survive contact with that market?
70stable
Idea survives; on-device constraint is a genuine moat where cloud can't reach.
1 evidence point backing this score
Trust Score
2 verified · 1 claimed · 0 contradicted
Every claim this founder made, traced to a source and rated by how well it holds up.
Verified — independently cross-checked against an outside sourceClaimed — founder's word only — not yet checkedInferred — reasoned from evidence, not stated outrightContradicted — checked, and it does not hold up
tractionLive pilot in 2 clinics
VerifiedChecked: Clinic confirmation email + on-device telemetry
“Deployed to 2 rural clinics, running offline” — Clinic pilot confirmation
productNovel 4-bit quantization for diagnostic CNNs
VerifiedChecked: Reproduced benchmark from public repo
“quant: 4-bit path, +2.1pp accuracy vs baseline at 1/6 size” — github.com/amara/kernel-health@7d20aa
teamSolo founder, self-taught
ClaimedLegibility
A great talker and a great builder can score the same on a deck. This strips out how well the pitch was told, and scores what was actually built.
Raw · the substance, stripped of delivery76
Polished · how good it sounded in the room59
Correction applied — the gap between substance and delivery is removed.
Corrected · what we actually act on76
SUBSTANCE RECOVERED FROM FLUENCY PENALTY