KINETICA AI
N-of-1 · longitudinal · open research

243 days of one body, turned into auditable clinical AI.

Pipeline · predictors · guarded agent · safety layer. End to end, open, reproducible.

WHAT KINETICA BUILDS

Clinical AI, built on real longitudinal physiology

Pipeline

One versioned longitudinal archive.

Wearable physiology and prospective symptom diaries land in a single reproducible dataset, L0 to L6. Every downstream model reads from the same contract.

Predictors

Idiographic N-of-1 models.

Trained on one patient’s deep physiology, not cohorts. Each targets a different clinical signal, validated leave-one-out with bootstrap CIs.

Architecture & safety

A guarded agent loop.

Clinical reasoning audited end to end, with a deterministic safety layer scoring every response against clinical boundaries before it reaches a clinician.

EVIDENCE

Results that hold up to inspection

243days · 71 features

The canonical longitudinal archive from a single subject — every model reads from it. The live ribbon above streams its most recent nights.

0.829 / 0.837AUC

Autonomic burden and symptom severity, leave-one-out, N-of-1. Bootstrap confidence intervals.

0.85retrieval acc.

Curated clinical RAG over 1,880 audited chunks, scored on a 20-question benchmark.

OPEN RESEARCH

Open research, verifiable systems

Every piece is public. The work below opens to code and a reproducible run, with claims anchored in peer-reviewed sources.

PIPELINE

Pipeline Polar & Symptoms

Polar exports and prospective symptom diaries flow into one clean, versioned longitudinal dataset — every level L0→L6 reproducible.

71 HRV features · deterministic nightly jobs

View pipeline
PREDICTOR

ANS Predictor

Multi-target models estimate symptom burden from nocturnal HRV and diary-linked physiology, on the same Polar pipeline.

AUC 0.829 autonomic · 0.837 severity · n=61

View predictor
PREDICTOR

Sleep Quality Predictor

Sleep quality treated as its own clinical signal — how nocturnal structure and autonomic patterns track perceived degradation and recovery.

AUC 0.77 · same physiological foundation

Explore sleep model
ANALYSIS

Cross-Predictor Convergence

Where two independent models agree: ANS and Sleep each selected nocturnal RMSSD as their top fatigue feature, on their own.

r=0.66 · 79% day-level agreement · n=42

Explore convergence
ABOUT

Engineeredfromphysics,biomechanicsandtenyearsofclinic

Kinetica AI is built by Alfonso Navarro. Physics at Universidad de Granada, with postgraduate work in biomechanics. Trained in osteopathy at UAB. Ten years of independent clinical practice in the Pyrenees: complex musculoskeletal and neuromechanical cases, high-performance athletes, mountain-sport injuries. Two years of acute COVID hospital care in Vielha during the pandemic.

He is also the patient. The system is engineered from real physiological uncertainty, not benchmark chasing — wearable monitoring, longitudinal symptom data and interpretable architectures for clinical AI.

1998Physics, Univ. Granada
2002Optometry training
2008Postgraduate in biomechanics
2016Independent clinical practice, Pyrenees
2020Acute COVID hospital care, Vielha (2 years)
2024Building clinical AI from own physiology
2025Kinetica AI
10 years clinicalOpen-source · MITMálaga · Remote
See open research →

Clinical AI consulting · Research-grade HRV analysis · AI model evaluation

Available for projects