What I can be brought in to do
Five engagement types, a technical stack chosen for longevity rather than fashion, and an honest list of the work I turn down.
32
Countries with teams trained on a single collection standard
1,000+
Bean varieties held in the PABRA database
30,000+
Investor records migrated without loss of service
90
Programme staff onboarded onto Microsoft 365
Five ways to engage
Data System Design & Build
End-to-end delivery of a research database, national registry or institutional platform: data model, migration, collection layer, curation workflow and handover documentation.
Best for
Institutions whose data is scattered across spreadsheets and individual machines.
Analytics & Dashboard Delivery
Harvesting data from disconnected internal systems via APIs, materialized views and flat files, then designing the reporting engine and visualizations decision makers actually use.
Best for
Ministries, universities and institutes with data they cannot see.
ICT Infrastructure Setup
Office networks from bare walls: LAN design, structured cabling, server rooms, Active Directory, VPN, firewall, antivirus, SSL, disaster recovery and equipment inventory.
Best for
New offices, satellite sites and organisations outgrowing ad-hoc IT.
Project Management & Business Analysis
Requirements, specifications, vendor SLAs, quality assurance and user acceptance testing on donor- and government-funded digital programmes.
Best for
Programmes with a vendor building the system and nobody holding the standard.
Capacity Building & Training
Structured training for field teams, programme staff and IT departments on data collection, collaboration tools, cyber security and system administration — with manuals and SOPs.
Best for
Projects where adoption, not the build, is the real risk.
Tools chosen for longevity
Databases & Data Management
Programming & Frameworks
Data Analysis & Visualization
Machine Learning & AI
Infrastructure & Networking
Cybersecurity
Sectors and representative work
| Sector | Experience | Representative Work |
|---|---|---|
| Agricultural research | 10+ years | PABRA database, ODK collection across 32 countries, PestDisplace, MaRLO, national dry bean adoption surveys |
| Government & e-services | 5+ years | NIFAMIS, UG Crafts, Ministry of Public Service dashboard, Works & Transport Statistical System, Webfarm |
| Health & health informatics | 4+ years | TeamUp data standards, CommCare case management rollout, mobile syndromic disease surveillance in abattoirs, MSc Health Informatics |
| Capital markets | 2+ years | Automated Trading System, IPO platform, biometric investor registration, SWIFT settlement, 30,000+ record migration |
| Higher education | 2+ years | Kyambogo University smart dashboard harvesting academics, HR, planning and finance data |
Why these choices, not others
Choose boring technology for systems that must outlive you
Oracle, SQL Server, PHP, ODK. Not fashionable choices; choices a successor institution can still hire for in ten years.
Design for the worst connection in the programme
If the satellite office on a VPN and the field officer with a tablet cannot use it, the platform is not deployed — it is demonstrated.
Standards before code
Metadata definitions and data security standards get agreed with implementing partners first. Retrofitting them into a live database costs ten times more.
The handover is the deliverable
Documentation, SOPs, inventories, SLAs and trained staff. A system nobody can administer is a liability with a login page.
What I do not do
I do not take on pure graphic design or brand identity work.
I do not build consumer mobile apps without a data or institutional backbone behind them.
I do not take engagements where training and handover are cut from scope to reduce cost.
I do not act as a sole vendor on projects that need a full delivery team — I will tell you and help you scope one.
Not sure which of these you need?
Describe the situation and I will tell you what the engagement should actually be — including if it is smaller than you think.