Skip to content
MLLabz

Capabilities

Software is the foundation. Intelligence is the direction.

Our capabilities exist to serve problems, not the other way round. We choose the tool after we understand the problem. See how we work →
Core today

Software Engineering

Platforms, applications, APIs, and systems that people and organisations use every day.

When it's the right tool
When work is manual, scattered, or slow, and a reliable system would fix it.
What we've built
  • a live education platform with payments by M-Pesa (CareerNext);
  • a property system with M-Pesa payment matching, USSD, offline field work, and strict separation of each organisation's data (LandlordPMS).
Developing

Data & Analytics

Collecting, cleaning, and analysing data so that it explains something and supports a decision.

When it's the right tool
When the information exists but nobody can see what it means.
What we've built
  • turned official KUCCPS PDFs and portal data into a clean course database (more than 14,000 course offerings), with several years of cutoff history and trend charts;
  • collection, arrears, profit and loss, and cash-flow reports in LandlordPMS.
Developing

Machine Learning

Systems that learn patterns from data, to predict, classify, or recommend.

When it's the right tool
When there's enough good data, and a prediction would change a decision. And only when the model beats a simple baseline.
Where we are
Learning, through Data Science study and Lab work. No model of ours is in production yet.
First planned
Estimating admission chances for CareerNext from several years of cutoffs, to replace the current rule-based trend bands.
Developing

Artificial Intelligence

Applications that understand language and images, assist people, and automate work safely.

When it's the right tool
When people need answers from complex information, or when a task is repetitive and language-heavy.
What we've built
  • an AI guide grounded in each student's own results, plus grade extraction from result-slip photos (CareerNext);
  • a staff assistant that always asks for confirmation before acting (LandlordPMS).

One system, three layers.

  1. Data collection
  2. Analytics
  3. Model
  4. API
  5. Interface
  6. Action
Software
The systems people and organisations interact with
Data
The systems that collect, organise, analyse, and explain information
Intelligence
The systems that learn, predict, recommend, automate, or assist

Intelligence on its own isn't useful. It becomes useful inside a system people already rely on. That's why we build the software and the data foundations first.

Careful with data. Careful with AI.

We use AI and data carefully, transparently, securely, and with respect for people. In practice:

  • AI that can act asks for confirmation first;
  • personal data is kept to what the problem needs;
  • every organisation's data stays separate;
  • we say plainly what a prediction can't promise.

Built with

TypeScript · Python · SQL · Next.js · Django · FastAPI · PostgreSQL

We choose technology because it fits the problem, not the other way round.

See the evidence →Bring a problem →