Case studies

Twelve principal projects, one body of work

Each project below is a full case study: the problem, my role, what was built, the architecture and the outcome. Modules and integrations are grouped under their parent project. Every individual project card is still in the Projects section below. Status labels separate what is in real-world use from what is still being built or designed.

Digital health & public health technology how the projects fit together

The common threads are interoperability (FHIR, HL7, APIs), data engineering, software architecture (.NET, PostgreSQL), analytics and AI, governance and security, and technical leadership.

Deployed in real-world use Pilot live in a pilot rollout In development product being built Prototype working code or MVP, validation pending Concept designed or specified Research interest

Select a case study to open it, or open them all.

01

Health-in-a-Box (HnB): Hospital Management Information System

CTO / Product Manager · Health-in-a-Box Solutions
Deployed

A modular hospital information system that digitises clinical, diagnostic, administrative and financial workflows in about 25 hospitals. It is my largest hospital-level digital health product.

  • ~25 hospitals
  • 4,500+ health workers trained
  • 25 federal institutions on the FMoH dashboard
  • On-prem · cloud · hybrid deployment
Read the full case studyClose case study
The problem

Disconnected systems, manual registers, duplicate patient records, fragmented laboratory and radiology workflows and poor revenue visibility, often in facilities with unreliable connectivity. These delay care and let revenue leak.

My role
  • Turned hospital operational requirements into product features and implementation plans; prioritised modules and coordinated engineering.
  • Designed and reviewed software architecture and database structures.
  • Coordinated hospital onboarding and deployment; integrated laboratory analysers, imaging and payment services.
  • Led training, adoption and troubleshooting; wrote deployment documentation for on-premises, cloud and hybrid environments.
  • Aligned the product with data-protection, security and interoperability requirements.
What was built
  • EMR: registration, encounters, history, diagnoses, prescriptions and longitudinal records.
  • LIS: orders, specimens, results and analyser interfaces that cut manual transcription.
  • RIS & PACS: radiology workflows with DICOM imaging linked to the record.
  • Pharmacy: prescribing, dispensing, inventory and drug-interaction alerts.
  • Specialist services: ART clinics, IVF, dialysis, cardiology, theatre, ECG, echocardiography and Holter monitoring.
  • Billing & revenue: bills, payments and balances, the foundation for CareWallet and SmartCare Card.
Products built on HnB
  • Informatics Dashboard (Product Manager): deployed at the Federal Ministry of Health. It aggregates visits, new patients, payer mix, wait times, doctor-to-patient ratio and treatment cost from 25 federal health institutions (15 FTHs, 7 FMCs, 3 SHs). See screenshot
  • FastCare: telemedicine app linked to HMIS records and appointments.
  • HIMIS: insurance workflows connecting providers, HMOs and state and public schemes.
Architecture & technology

Configurable workflows, role-based permissions and secure record access across hospitals with different needs. Offline synchronisation with conflict resolution, retries, audit and duplicate-transaction controls. Interoperability direction: HL7, FHIR and terminology mapping.

  • C# / .NET
  • ASP.NET MVC
  • EF Core
  • SQL Server
  • PostgreSQL
  • JavaScript
  • React
  • Bootstrap
  • SignalR
  • Azure
  • Git
  • HL7 / FHIR
  • DICOM
  • ASTM
Outcome

One platform connecting clinical care, diagnostics, hospital operations and revenue management across about 25 hospitals, with a ministry-level view through the FMoH dashboard.

Product site
02

SORMAS Optimisation & Re-engineering: national disease surveillance

Health Informatics Expert / Technical Lead · IHVN with NCDC · U.S. CDC-funded SECURE-Nigeria
Deployed

Modernisation of Nigeria's electronic surveillance and outbreak-management platform: a version upgrade, re-engineered disease forms, performance work and an NCDC team trained to run it. National go-live was on 27 February 2026, across all 774 LGAs.

  • v1.82 → v1.101 upgrade
  • 18 disease-specific forms
  • ~25% faster form loading
  • 15+ min → <3 min large downloads
Read the full case studyClose case study
The problem

A national system covering many diseases, large record volumes and users at federal, state and LGA level. It had fallen 19 versions behind, its forms loaded slowly, disease workflows were inconsistent, and maintenance depended on outside developers.

My role

Technical lead for software optimisation, form re-engineering, integration, problem-solving and workforce capacity strengthening (implementation March–September 2025).

What was done
  • Version upgrade from 1.82.0 to 1.101.0, covering compatibility, data integrity, regression testing and deployment procedures.
  • Form re-engineering: replaced ReactJS case investigation forms with JavaServer Faces (JSF) implementations for performance and fit.
  • 18 disease-specific forms with validation rules matching national paper forms.
  • Laboratory results linked to cases by EPID number, a faster download manager and a contingency line-list upload.
  • Multi-level user validation (a 23-item instrument) to prioritise 40+ issues before go-live.
Capacity & governance

Trained 14 of 15 targeted NCDC Public Health Informatics personnel to maintain, develop and troubleshoot the platform. Co-designed the governance framework NCDC adopted: Git version control, change requests, and dev → staging → production.

  • Java
  • JSF
  • ReactJS
  • PostgreSQL
  • Regression testing
  • Stata
  • IDSR
Context

The wider national deployment serves about 5,000 users across 1,000+ facilities. Those figures describe the whole national system, not this project alone.

See before & after SORMAS case study (IRCE)
03

DAVT: Data Analytics and Visualization Tool (SECURE-Nigeria)

Architect / Technical Lead · IHVN & NCDC
Deployed

NCDC's surveillance analytics layer, which turns SORMAS and other data into dashboards, the Weekly Epidemiological Report, alerts and decisions. It has been used in all 36 states and the FCT since January 2024.

  • 6 → 37 states + FCT
  • 10 analysis modules
  • 5 connected data sources
  • 5–6 wks → real time weekly report
Read the full case studyClose case study
The problem

Surveillance data was collected but hard to turn into evidence. Reports were compiled by hand and late, there was no single outbreak view, decisions were recorded manually, and there was no automatic cross-check of SORMAS against other sources.

Objectives
  • Integrate surveillance sources and reliably extract and transform records.
  • Provide interactive dashboards and trend analysis by place and time.
  • Make routine reporting efficient and consistent.
  • Strengthen traceability with audit trails and controlled access.
  • Build an extensible base for advanced statistics and prediction.
What was built
  • Standalone SORMAS data-pull service that separates ingestion from the dashboards.
  • Dashboards & analytics: summary, aggregate and case-based views, geospatial and temporal analysis, and a live EOC wallboard.
  • Weekly Epidemiological Report generated from live data.
  • Threshold alerts and a decision tracker from alert to resolution.
  • Triangulation: SORMAS, IDSR-002, state line lists, DHIS2 (NPHCDA) and MedSafety (NAFDAC).
  • Audit trail recording system activity for accountability.
Data engineering

Extraction schedules, pagination, schema compatibility, failed requests, duplicate prevention, separating missing values from zero counts, and traceability from results back to source records.

Next: built & specified
  • Built Digital WER workflow, automatic situation reports, the National IDSR Compliance Index and advanced analytics, awaiting NCDC adoption.
  • Specified DAVT Analytics Studio: hybrid R/Python execution, row-level access by default, multi-level geography, time-series, spatiotemporal and regression analysis, and reproducible exports for high-volume data.
  • .NET
  • JavaScript
  • SQL
  • Power BI
  • SORMAS API
  • DHIS2
  • GIS
  • R / Python
See screenshots DAVT case study (IRCE) SDQM data-quality case study
04

National Hospital Abuja LIMS: laboratory digitisation & analyser integration

Scientific Officer · EMR/LIMS Administrator & Quality Officer
Deployed

Digitised laboratory workflows across the hospital's four major laboratories and connected the analysers directly to the information systems.

  • 4 major laboratories
  • Abbott · Roche Cobas · Sysmex · Selectra interfaced
  • Bidirectional LIS/EMR links
Read the full case studyClose case study
Workflow digitised
  1. Test ordering by clinicians
  2. Specimen registration and tracking
  3. Routing to the right section or analyser
  4. Result capture, review and validation
  5. Reporting to clinicians
  6. Audit and traceability
Analyser middleware

Built the integration layer between instruments and the LIS, including MasterLISMiddleware for Abbott Architect analysers:

  • Mapping instrument test codes, units, reference ranges and specimen IDs to the test catalogue.
  • Order-to-result association and duplicate-import prevention.
  • Handling invalid or unexpected messages, with error logging and message recovery.
  • Preserving result provenance and review status.
Patient safety & integration

Every result stays tied to the right patient, encounter, specimen and test, with safeguards against ID mismatches, unit errors, wrong mappings and overwriting validated results. Imported results are kept separate from results authorised by laboratory staff. The LIS is linked to the EMR and PACS/DICOM.

  • ASTM
  • HL7
  • LOINC
  • FHIR
  • Middleware
  • Quality management
Analysers I've worked with
05

NHEx: National Health Information Exchange & FHIR architecture

Chief Architect / Development Lead
Pilot

A national interoperability layer that lets hospitals, laboratories, telemedicine, surveillance and insurance systems exchange authorised health information without replacing their own software.

  • 3 national registries
  • 24 teaching hospitals targeted in Phase 1
  • FHIR on .NET + PostgreSQL
Read the full case studyClose case study
The problem

Many systems with different data models, terminology, authentication and interfaces lead to fragmented histories, repeated tests, inconsistent identifiers and manual transfers.

Ecosystem

Health-in-a-Box, external EMRs, LISphere and laboratory middleware, FastCare telemedicine, SORMAS, insurance, national blood management and pharmaceutical supply (Medex).

Identity & consent

National Health ID anchored to the NIN, with alternative identifiers, demographic matching and correction of wrong merges. Consent by card, OTP or app; break-glass access for emergencies; full audit.

FHIR resource model Proposed mapping
AreaResources
IdentityPatient, RelatedPerson
EncountersEncounter, Condition
DiagnosticsServiceRequest, Observation, DiagnosticReport
MedicationMedicationRequest, MedicationStatement
LaboratorySpecimen, Observation, DiagnosticReport
ProvidersPractitioner, PractitionerRole, Organization
TrustConsent, AuditEvent, Provenance
Architecture Proposed
  1. Participating systems
  2. API gateway: FHIR endpoints, authentication, authorisation, validation, rate limits
  3. FHIR service: validation, terminology, patient matching, events
  4. Relational + JSONB persistence
  5. Search index and event/analytics layer

Event-driven processing with CQRS where it adds clear value. Security: encryption, minimum-necessary disclosure and NDPA compliance.

NHEx platform site
06

LISphere, LiSphere Connect & Quanta™ EQAS: laboratory software suite

Creator · CEO, efacilities Systems & Services
In development

Three complementary products: a laboratory information system (LISphere), instrument connectivity (LiSphere Connect) and external quality assessment (Quanta™ EQAS).

  • 460+ instrument models
  • 160+ manufacturers
  • 356 ms query round trip
  • Abuja · Waterloo
Read the full case studyClose case study
LISphere: laboratory information system

Laboratory configuration (departments, test catalogues, specimen types, reference ranges and users); orders, accessioning and traceability; result validation with units, flags and provenance; turnaround and workload reporting; quality control and audit.

  • Next.js
  • .NET APIs
  • PostgreSQL
  • FHIR
  • LOINC
LiSphere Connect: instrument connectivity Product
  • Instrument registration, protocol adapters, parsing and validation.
  • Test-code and unit mapping, specimen and order correlation.
  • Acknowledgement, duplicate detection, retry and error queues.
  • Interface audit logs, health checks and 24/7 monitoring.
  • Speaks ASTM, HL7 v2, serial, TCP/IP, file drop, cloud gateway and FHIR R4.
Quanta™ EQAS: external quality assessment Concept

Participant registration, assessment rounds, sample distribution, result submission, statistical analysis, performance scoring, outlier detection and corrective-action tracking.

Analysers supported LiSphere brand profile
07

CareWallet: healthcare fintech, payment orchestration & revenue

Product & Systems Design Lead
In development

A payment layer connecting patients, hospitals, EMRs and payment providers. It links every transaction to the patient, facility and service, and keeps the hospital in control of its own finances.

  • 3 wallet layers
  • Any EMR, not only HnB
  • Multi-processor strategy
Read the full case studyClose case study
The problem

Deposits, charges, cash, transfers, outstanding bills, refunds and settlement often sit in disconnected processes. The result is delayed revenue recognition, mismatched balances and transactions that are hard to trace.

Wallet ecosystem
  • Patient e-wallet: deposits, payments, refunds and balance.
  • Hospital settlement wallet: maps collected funds to the facility.
  • TSA or designated account: onward settlement under each hospital's approved arrangement.
Ledger design

The HnB/EMR keeps the patient-level ledger as the source of truth. It records funding, charges, allocations, reversals, refunds and settlement advice, using idempotency keys, an immutable history and explicit reversals. An API call is never treated as proof that a payment happened.

Integrations
  • Moniepoint: reserved-account (pool and settlement with daily advice) and wallet models; webhooks, statements and settlement APIs.
  • Parkway: virtual and static wallets, POS SDK.
  • Wema Bank / ALATPay: co-branded Verve cards and payment integration.
Beyond HnB

Deployable at facilities on other EMRs through APIs for patient identification, billing references, payment initiation, status, refunds and reconciliation.

  • Payment orchestration
  • Ledger design
  • Webhooks
  • Reconciliation
  • Settlement
  • Commercial strategy
08

SabiPHC / SmartPHC: offline-first digital primary healthcare

Creator / Project Lead
Concept

An offline-first, AI-assisted platform for primary healthcare in Nigeria, built around Standing Orders and IDSR workflows for JCHEWs, CHEWs and CHOs. It syncs to a central FHIR repository when connectivity returns.

  • Offline-first by design
  • Multi-cadre workflows
  • FHIR sync to the centre
Read the full case studyClose case study
Functional scope
  • Patient registration and longitudinal records, including patients with no national ID.
  • Encounters, observations, diagnoses, treatment, referral and follow-up.
  • Maternal and child health, immunisation and community outreach, built against defined programme requirements.
  • Reporting and supervision under access controls.
Offline-first engineering
  • Protected local storage, sync queue and retries.
  • Locally unique IDs and idempotent server processing.
  • Conflict detection that never lets last-write-wins overwrite clinically significant data.
  • Recovery after interruption, plus encryption and device controls.
Interoperability & responsible AI

Planned REST APIs to DHIS2, SORMAS and other EMRs, keeping clinical records, aggregate indicators and surveillance notifications distinct. AI is limited to explicit, validated use cases with human review, and never silently replaces clinical judgement.

  • Offline sync
  • HL7 FHIR
  • Danger-sign rules
  • Responsible AI
09

PlasBench, PlasmiCord & PlasmidTrace: plasmid-AMR bioinformatics

Bioinformatics Analyst & Developer · Creator
Prototype

A connected research portfolio on how antimicrobial-resistance plasmids spread. It covers benchmarking plasmid reconstruction, mapping plasmid relatedness, and attaching honest confidence to transmission signals.

  • 3 linked tools
  • 4 reconstruction tools benchmarked
  • Reproducible and provenance-tracked
Read the full case studyClose case study
Why it matters

Plasmids can carry resistance genes between bacteria. Finding the same gene in two isolates does not prove transmission. That needs plasmid similarity, genomic context, bacterial relatedness, metadata and careful interpretation.

The three tools
  • PlasBench: benchmarks MOB-recon, Platon, plasmidSPAdes and gplas against truth sets (precision, recall, F1, failure modes). GitHub
  • PlasmiCord: pairwise plasmid distances, clustering and isolate-sharing networks to find candidate horizontal spread. GitHub
  • PlasmidTrace: links reconstruction reliability with sharing evidence to produce confidence scores; needs external validation.
Shared pipeline
  1. Sequencing data & metadata (FASTQ, FASTA, assemblies)
  2. QC & genome processing
  3. Annotation & resistance profiling (Prokka, ResFinder, CARD)
  4. Plasmid comparison, clustering & phylogeny
  5. Evidence-based interpretation & reproducible reports
Research quality

Records software and database versions, parameters, input checksums, QC metrics and output provenance. Keeps observed data separate from inferred relationships, documents uncertainty, and never treats a database match as proof of resistance or transmission.

  • Python
  • Ubuntu / Bash
  • Prokka
  • ResFinder
  • CARD
  • Phylogenetics
  • Genomic epidemiology
10

REACH: Artificial Intelligence to Accelerate Humanitarian Giving

Creator / Project Lead
Beta MVP

An AI-enabled platform to help donors and humanitarian organisations direct resources to credible needs and to show where support goes. The poster was accepted at IRCE ISS2026 (No. ISS26_054).

  • Beta MVP built
  • Poster accepted, ISS2026
  • GDPR / NDPA-aware
Read the full case studyClose case study
The problem

Humanitarian information is fragmented across organisations and reports. Donors struggle to judge credible needs, and organisations struggle to present their needs in a comparable form.

Product framework
  • Structured needs intake and organisation.
  • AI-assisted categorisation, document extraction and matching, transparent and human-reviewed.
  • Donor and organisation engagement.
  • Impact reporting and accountability.
Responsible AI

Data minimisation, consent, access control and encryption; human oversight of recommendations; bias review across geography, language and connectivity. No automated judgement of who deserves support.

  • NLP
  • Semantic search
  • Recommendation
  • Responsible AI
11

CODA: Clinical Ontology, Discovery and Assistance

Creator · Product & Technical Lead
In development

A clinical terminology intelligence platform that searches, validates and explains clinical codes across terminologies. Its core is deterministic and versioned; AI assists, but is never the source of truth for a code.

  • ICD-11 · LOINC · RxNorm · SNOMED CT and UMLS sources
  • 6 public SDKs
  • .NET 10 + React workbench
  • Audited coding decisions
Read the full case studyClose case study
The problem

Clinical data is only as good as its codes. Teams code by hand across ICD, LOINC, RxNorm and SNOMED CT with no shared validation, no audit of why a code was chosen, and no safe way to manage local codes or mappings.

What CODA does
  • Search, look up, validate and explain codes across terminologies, served live from their publishers (WHO ICD-11, LOINC, RxNorm and other UMLS sources) under each organisation's licences.
  • Every selection and review is recorded as an audited decision.
  • Local dictionaries and mappings with four-eyes review, plus value sets.
  • An AI phrase assistant that suggests candidates; deterministic rules decide.
  • An EMR sandbox and embeddable components to drop terminology search into any clinical app.
Architecture & engineering
  • API: .NET 10 modular monolith, PostgreSQL, OIDC/JWT (authorisation code + PKCE), role-based access and a full audit trail.
  • Web: React + TypeScript coding workbench and reusable components.
  • Resilience: a local copy of every terminology, refreshed monthly, answers when a live source is down.
  • Interoperability: FHIR adapter; SNOMED CT through an evaluated terminology engine.
  • Quality: CI with build, tests, vulnerability scans, Playwright end-to-end journeys, load tests and observability.
SDKs for integrators

Thin, dependency-light clients with one shared contract in C#/.NET, TypeScript/JavaScript, Python, Java, Go and Rust, released with verified checksums.

  • .NET 10
  • PostgreSQL
  • React
  • TypeScript
  • OIDC
  • FHIR
  • ICD-11
  • LOINC
  • RxNorm
  • SNOMED CT
CODA SDKs on GitHub
12

BodyChart: HMIS Medical Body Charting Platform

Creator · Personal project
Prototype

An enterprise patient body-charting engine. Clinicians document findings directly on an interactive 3D/2D anatomical model instead of paper diagrams, across nursing, emergency, surgery, dermatology, wound care, burns, pain and obstetrics.

  • 3D + 2D print-safe charting
  • 8 body habitus variants
  • SNOMED-coded body regions
  • Signed revisions & amendments
Read the full case studyClose case study
Clinical capabilities
  • Smart markers, freehand drawing, region shading and severity heatmaps.
  • Burn surface area (adult rule of nines, paediatric tables, Lund-Browder) with Parkland fluid calculation.
  • Wound tracking with healing trends; pain mapping with structured OPQRST capture.
  • Validated instruments with shown workings: Braden, PUSH, LANSS, FLACC, ABCDE.
  • Dermatology lesion mapping and a dental odontogram in the same record.
Record integrity
  • Named save points with versioned sign-off, amendments and timeline playback.
  • Every marking stores its anatomical region and a 2D projection, so print and low-end devices never need 3D.
  • Offline drafts in the browser for recovery after crashes or network loss.
  • Local decision support with rationale, and no patient data leaving the device.
Architecture

A reusable React engine (@bodychart/core) that embeds in any EMR, plus a single-page clinical workstation. All charting runs in the browser; the backend is specified as a document-based API contract.

  • React 18
  • TypeScript
  • Three.js
  • React Three Fiber
  • Zustand
  • MUI
  • IndexedDB
  • Zod
  • Vitest
Projects & innovation

What I've built and what I'm building

Systems in real-world use, products I lead, and research prototypes. Each card shows its status, so it's clear what is live and what is still being developed.

Deployed in real-world use Pilot live in a pilot rollout Product product / programme work Prototype working code, validation pending Concept design stage Research interest

Showing all projects

DAVT Next-Generation Modules

Designer / Builder
Prototype

Four modules built and working inside DAVT, awaiting NCDC review and adoption. They need no new data collection.

MoreLess
  • Digital WER: auto-drafted every Monday, with an approval workflow and locked, versioned editions.
  • Situation reports: generated for any disease straight from SORMAS case records.
  • National IDSR Compliance Index: a monthly 0–100 score for the nation, states and LGAs, including 7-1-7 timeliness.
  • Advanced analytics: seven transmission models, R estimated four ways, forecasts and a "what if" simulator.
  • Epidemic modelling
  • Forecasting
  • Workflow automation
  • 7-1-7

Surveillance Data Quality Management

Implementation Lead
Deployed

Monthly data-quality reviews at NCDC and in states, run on DAVT's concurrent-trends view, finding the causes of reporting gaps and fixing them.

MoreLess
  • Pilot in Ondo, Sokoto, Ebonyi and the FCT (Oct 2023 – Sep 2024), later expanded to seven states.
  • Four measures (completeness against line lists, key fields, timeliness, LGA reporting) across 10 priority diseases.
  • Yellow Fever paper-to-SORMAS ratio: 1.64 (2022) → 0.99 (2025).
  • Data quality
  • Concordance
  • Root-cause analysis
SDQM case study

National Hospital Abuja LIS

Designer / Developer / Administrator
Deployed

A Laboratory Information System for the hospital's four major laboratories, with bidirectional analyser interfaces.

  • Abbott
  • Roche Cobas
  • Selectra
  • Sysmex
  • ASTM
  • PACS
Analysers I've worked with

FastCare Telemedicine

Product & Integration Lead
Product

A mobile telemedicine app linked to HMIS records, appointments and continuity-of-care workflows.

  • Mobile
  • Telemedicine
  • HMIS integration

CareWallet

Product & Systems Design
Product

Healthcare payments and patient-wallet workflows integrated with hospital settlement and HMIS billing.

  • Payments
  • Webhooks
  • Reconciliation

HIMIS Insurance Workflows

Requirements & Product Contribution
Product

A health-insurance system linked to HMIS, connecting providers, HMOs, state schemes and public programmes on eligibility, billing and claims.

  • Claims
  • BHCPF
  • NEMSAS

BodyChart: HMIS Medical Body Chart

Creator · Personal project
Prototype

A 3D/2D anatomical body-charting engine for EMRs: wounds, burns (TBSA and Parkland), pain, lesions and validated clinical scores, with signed revisions.

  • React
  • Three.js
  • TypeScript
  • SNOMED regions
Read the case study

PDX: Portable Health Records

Creator / Research Developer
Prototype

Patient-controlled, offline-first record portability across different EMRs, with identity, consent, provenance and audit built in.

  • FHIR
  • openEHR
  • Offline-first

PlasBench

Creator / Research Developer
Prototype

A reproducible benchmark of plasmid reconstruction tools against trusted plasmid/chromosome truth sets.

MoreLess
  • Benchmarks MOB-recon, Platon, plasmidSPAdes and gplas; normalises outputs and evaluates precision, recall, F1 and failure modes.
  • Planned stratification by species, plasmid size, replicon, coverage, copy number, AMR genes and assembly quality.
  • Python
  • Bash
  • Benchmarking
  • AMR
plasbench

PlasmiCord

Creator / Research Developer
Prototype

A plasmid-aware framework for relatedness, sharing networks and chromosome–plasmid discordance.

MoreLess
  • Pairwise plasmid distances, clustering and isolate-sharing networks reveal candidate horizontal spread that chromosome-only surveillance misses.
  • Provenance and validated thresholds guard against unsupported cross-site inferences.
  • Python
  • Networks
  • Comparative genomics
plasmicord

PlasmidTrace

Creator / Research Developer
Prototype

An analytical engine that links reconstruction reliability (PlasBench) with plasmid-sharing evidence (PlasmiCord) and scores confidence.

MoreLess
  • Runs separate calibration and surveillance tracks; similarity alone is never treated as proof of transmission.
  • Outputs indices, clusters, confidence scores, networks and discordance signals for analyst review. Needs real-data and external validation before decision-grade use.
  • Confidence scoring
  • AMR linkage

Myconductor TB

Genomic research
Research interest

Reproducible analysis of archived M. tuberculosis complex FASTQ data alongside DST/MIC phenotypes.

  • MTBC
  • FASTQ
  • DST/MIC
myconductor

SabiPHC

Creator / Project Lead
Concept

Offline-first, AI-assisted decision support for JCHEWs, CHEWs and CHOs, built on Nigeria's Standing Orders and IDSR workflows.

MoreLess
  • Deterministic danger-sign rules, triage, maternal and child health, referral and reporting.
  • Mandatory health-worker review and responsible-AI controls; staged validation required before real-patient use.
  • Offline-first
  • Responsible AI
  • PHC

RegX Research Studio

Creator / Project Lead
Concept

Field-research data capture: voice-to-structured data, OCR extraction, form generation and offline-first collection.

  • Voice-to-data
  • OCR
  • Planned: SORMAS, DHIS2, Kobo, REDCap

REACH Analytical Studio

Creator / Project Lead
Concept

A research and analytics environment for public health, disease/biological modelling and AI-assisted analysis with human oversight.

  • Modelling
  • Responsible AI
  • Poster accepted: IRCE ISS2026