Careers

    Build the systems institutions actually run on

    We are a small engineering team behind ERP, MIS and climate-intelligence platforms used by governments, UN agencies and development banks across six African countries. One role is open right now.

    Apply for the AI/ML Engineer roleEmail us

    Open role

    AI/ML Engineer

    Build and ship the models behind our drought early-warning, beneficiary targeting and programme analytics platforms — systems that government and UN teams act on, in production, across six countries.

    • JKUAT Towers, Nairobi, Kenya
    • Hybrid — part of each week in Nairobi
    • Full-time
    • Applications close 28 September 2026

    This is a hands-on engineering role, not a research post. The models you build go into systems that district officers, ministry analysts and humanitarian responders use to make decisions with real consequences — so they have to be accurate, explainable, and dependable on the infrastructure our clients actually have.

    You will work alongside the platform and delivery teams on Early Warning Information Systems running in production today, and you will be in the room when the institution reviews what the model recommended and why.

    What you will do

    • Design, train and evaluate models for drought and flood risk forecasting, beneficiary targeting, anomaly detection and forecasting across programme data.
    • Take models to production: packaging, serving, monitoring, retraining schedules, and the fallback behaviour when a data source goes quiet.
    • Build the feature pipelines — fusing satellite, IoT sensor, survey and administrative data, much of it incomplete and irregularly delivered.
    • Make every consequential output explainable. If a model flags a sub-county or excludes a household, we must be able to show why to an auditor.
    • Work with the client's own analysts so they can operate and interrogate the system after handover.
    • Contribute to proposals and technical assessments when a programme is being scoped.

    What we need from you

    • 3+ years building and deploying machine learning systems, at least one of which went into production and stayed there.
    • Strong Python, and fluency with the modern ML stack — scikit-learn, PyTorch or TensorFlow, pandas, and a workflow orchestrator.
    • Solid grounding in statistics and model evaluation: you can tell a genuinely useful model from one that has learned the test set.
    • SQL and comfort working with messy, real-world relational and geospatial data.
    • Able to explain a model's behaviour and its limits to a non-technical audience — a programme lead, a director, an auditor.
    • Based in Nairobi or able to relocate. The role is hybrid: part of each week in the Nairobi office, the rest wherever you work best.

    Nice to have

    • Geospatial and remote sensing experience — Earth observation data, raster pipelines, GDAL/rasterio, Google Earth Engine.
    • Time-series forecasting at operational cadence.
    • MLOps tooling: MLflow, Docker, CI for models, drift monitoring.
    • Prior work on public-sector, humanitarian or development programmes, and familiarity with what a donor audit asks for.
    • Experience with responsible AI practice — bias assessment, human-in-the-loop design, model documentation.

    What we offer

    • Live systems with real users, not proofs of concept that die in a notebook.
    • Direct exposure to government, UN and development-bank counterparts.
    • A team that treats auditability and explainability as engineering requirements, not paperwork.
    • Competitive salary, reviewed against the Nairobi market, with a structured progression path.

    Apply for this role

    Applications go straight to careers@rosewillbome.com. Fields marked * are required.

    Share a link rather than a file, and please check it is viewable by anyone with the link.

    This is the part we read most closely. Specifics beat adjectives.

    We use what you send here only to assess this application. See our privacy policy.

    Why here

    What the work is actually like

    Systems, not demos

    Everything we build goes into production and stays there. You will not spend a year on something that never leaves a notebook.

    Accountable by design

    Auditability and explainability are engineering requirements here, because our clients answer to parliaments, boards and auditors.

    In the room

    Engineers sit with the institutions we serve. You will hear directly what a district officer needs, not a summary of it.

    Nothing open that fits you? We still read speculative applications. Send a CV and a note about what you want to work on to careers@rosewillbome.com.

    Not looking, but know someone who is?

    Referrals from people who have worked with us carry real weight. Send them the role on WhatsApp, or email us their name.

    Refer someone on WhatsAppEmail us their name
    • Response within 24 hours
    • NDA available on request
    • No-obligation discovery call