HAKILIX | Musah Shaibu - Principal Investigator

// 02_TECHNICAL_ARCHITECTURE

Proprietary Sensor Fusion.

Standard AI predicts correlation. Hakilix infers causation. We fuse 4D mmWave Radar with Radiometric Thermal Imaging.

Spiking Neural Networks

Utilizing Leaky Integrate-and-Fire (LIF) models to process temporal data spikes with biological energy efficiency.

EQ: dV/dt = -(V - Vrest) + Input

Causal Discovery

Structural Causal Models (SCM) allow the system to reason counterfactually: "Did they fall, or lie down intentionally?"

METHOD: DIRECTED ACYCLIC GRAPH

Homomorphic Encryption

Federated Learning enables the global model to learn from encrypted gradients without ever exposing raw patient data.

STATUS: GDPR/CQC COMPLIANT

// 03_THEORETICAL_MODEL

Minimizing Variational Free Energy.

The system acts as a Markov Blanket, separating internal states from external sensory states. Hakilix minimizes "Surprisal" ($F$) to maintain homeostatic equilibrium.

EQ 3.1: FREE ENERGY PRINCIPLE

F = D KL [q(s) || p(s|o)] - ln p(o)

neuromorphic_core.py

# Leaky Integrate-and-Fire (LIF) Neuron Model

class RadarSpikeEncoder(nn.Module):

def forward(self, radar_frame):

""" Converts continuous signals to temporal spikes """

self.mem = self.mem * 0.9 + radar_frame

spike = (self.mem > self.threshold).float()

self.mem = self.mem - (spike * self.threshold)

return spike

// 03_SYSTEM_ARCHITECTURE

The Edge-Cloud Continuum.

A distributed, privacy-preserving topology designed for high-latency environments.

Local Active Inference

The device acts as a Markov Blanket. It minimizes variational free energy locally, allowing offline fall detection.

Federated Aggregation

Only encrypted gradients are transmitted. The global model learns without accessing private user data.

// 04_STRATEGIC_IMPACT

Solving the £18bn Ageing Crisis.

Hakilix is critical national infrastructure for the Silver Economy. By shifting care from reactive (hospital) to proactive (home), we unlock massive economic value.

0%

Reduction in Acute Admissions

£0k

Savings per QALY (NHS)

0M+

Target UK Demographic

0%

Privacy Compromise

// 05_EXECUTION_PROTOCOL

36-Month R&D Horizon.

Hardware Engineering

PCB Design & Sensor Integration

01

MONTHS 0-9

Integration of 60GHz FMCW Radar (TI IWR6843) with 32x32 Thermal Array.

DELIVERABLE: TRL 4 PROTOTYPE

MONTHS 6-18

Training SNN on synthetic data. Implementing "Cross-Modal Hallucination".

DELIVERABLE: >90% ACCURACY

02

Neuromorphic Modeling

Field Validation

03

MONTHS 18-36

Deployment in 220 volunteer homes. Security audit of Homomorphic Encryption.

DELIVERABLE: CLINICAL TRIAL

// 06_PRINCIPAL_INVESTIGATOR

Musah Shaibu

AI RESEARCHER & SYSTEMS ENGINEER

Specialising in neuromorphic machine intelligence, causal modelling, and privacy-preserving computational frameworks. Proficient in translating mathematical frameworks (SNNs, Active Inference) into deployable real-time systems for healthcare.

RESEARCH FOCUS

Neuromorphic Computing Active Inference Federated Learning Causal AI

EDUCATION

Advanced Computer Science
University of Liverpool
Computer Science
Catholic University College of Ghana

SELECTED PROJECTS

> HAKILIX: Autonomous Bio-Digital Twinning via Neuromorphic Active Inference.
> IoT-Enabled Pediatric Cardiac Anomaly Detection: A Multi-Modal Sensor Fusion Approach for Early Myocarditis Diagnosis.
> Sustainable Energy Analytics: Time-series forecasting for energy optimisation.

Musah Shaibu

PRINCIPAL INVESTIGATOR

Liverpool, United Kingdom

PROGRAMMING Python, C++, Java, SQL

AI/ML STACK TensorFlow, PyTorch, SNNs

CLOUD/DATA AWS, Docker, Kafka

MATHEMATICS Optimization, Bayesian Inf.

EMAIL +44 7708...