About us
Precision Pricing, South African Scale
Mutuka Automotive was founded to bring machine-learning rigour to South Africa's second-hand vehicle market. Where traditional appraisers rely on gut feel, we rely on 500 Random Forest trees trained on 201 real-market vehicles.
Every valuation is reproducible, auditable and explainable — three things a human appraiser has never once been.
Our Machine Learning Approach
We train on 10 vehicle features: engine displacement, horsepower, cylinders, city and highway fuel economy, kerb weight and bore ratio. Brand-adjusted correction factors are applied post-prediction for the SA market.
The result: an R² of 0.87 on our test set — well above the linear regression baseline of 0.71.
South Africa's Automotive Valuation Experts
Meet the team
A small, data-obsessed team keeping the model honest.
Sipho Nkosi
Chief Everything & Professional Storytelling Officer
Sipho founded Mutuka after spending three years watching dealers overpay and underpay for vehicles in equal measure. He is the reason we have a business plan and also the reason the business plan changes every six weeks.
Lerato Dlamini
Chief Money Wrangler
Lerato keeps the spreadsheets and the investors equally happy. Her background is in quantitative finance, which means she immediately understood what a Random Forest was and why it was better than asking a cousin. She has strong opinions about budget forecasts and even stronger opinions about coffee.
Yusuf Karriem
Head of Organised Chaos (Product)
Yusuf holds the vision for the Mutuka platform and communicates it to engineering via a combination of Figma mockups, voice notes and interpretive whiteboard diagrams. His product decisions are bold, occasionally correct, and always confidently explained.
Naledi Mokoena
Chief Bug Squasher & Coffee Drinker
Naledi wrote the prediction pipeline, the API, the dashboard and the deployment scripts. She also maintains the espresso machine. If it runs and it is fast, Naledi built it. If it is slow, Naledi knows about it and is already fixing it.