Rigorous quantitative analysis, machine learning, and econometric modelling — applied to decisions that matter.
Ferra Solutions started in 1997 as "MG's quantitative and investment services" with a focus on delivering rigorous quantitative analysis to the investment community.
Since its inception, the company has developed and applied models to support strategic and management decision-making — expanding into a substantial body of analysis for policymakers, including the National Treasury and the central bank.
Ferra Solutions began integrating machine learning techniques into its work in 2004 — long before the term became mainstream. These approaches have since been combined with established econometric and time series methods, enabling more flexible and robust solutions across a wide range of applications.
Alongside analytical work, Ferra Solutions has developed specialised modelling tools, including a platform that has supported the modelling and reporting of monetary policy decisions since 2012.
Our work has been recognised across leading platforms and international collaborations, reflecting both depth of expertise and a consistent ability to perform in highly competitive environments.
A challenge to determine whether AI could reconstruct the shuffled pages of a famously difficult 1934 literary puzzle and solve embedded murder mysteries. Our model achieved first place, demonstrating advanced reasoning and pattern recognition.
Challenges ranging from mathematical reasoning to real-time tracking and prediction tasks. We use Kaggle to benchmark methods and learn from leading practitioners globally — earning medals across diverse domains.
A long-standing engagement with Africa's premier data science competition platform — reflecting our commitment to the African data science ecosystem and our ability to deliver consistent, high-quality results across a wide range of challenges.
Ferra Solutions provides a diverse range of modelling related services.
If you are looking for a full pipeline to predict or analyse something of interest to you, or for technical assistance to apply the latest machine learning techniques to your data, or perhaps require a review of your model suite, then you have come to the right place.
We recognise that sophisticated models alone don't solve problems. Real value comes from applying them thoughtfully — delivering insights that are both meaningful and actionable.
A comprehensive range of modelling services tailored to your needs:
From central banks to commodity traders, our models have supported decisions that matter — across markets, governments, and institutions for nearly three decades.
Pioneered neural network-based algorithmic trading in 2004 — built from scratch before the technology was mainstream. Optimised on a locally-constructed compute cluster augmented with AWS EC2. Traded SAFEX instruments including ALSI and maize contracts.
Integrated cost-at-risk and bond calculator supported by a South African econometric core with US, Japan, and UK satellite models. Monte Carlo simulations used to optimise the government's funding strategy across multiple economic scenarios.
Integrated big bank income statement and balance sheet models with internal econometric models. Multiple univariate models estimated default rates per asset category per bank. Stress-testing scenarios developed — validated in practice during Nene-gate.
Economic indicator models using PESTLE-classified indicators analysed the external environment. A default rate model assessed internal risks. Quantitative inputs sketched emerging trends and shaped the bank's strategic preparation for an uncertain future.
A COTS solution delivered over 15 years with advanced time series functionality, univariate and multivariate econometric model support, and press kit publication capability. Developed in Java with proprietary high-performance NoSQL storage. IP acquired by the bank.
A suite of machine learning models delivering robust near-term spot price forecasts. A growing collection of model outputs used to calibrate assumptions and provide reliable starting points for downstream modelling processes.
Deployed shared Python and JupyterLab environments across an array of Linux servers, transforming the bank's modelling capability. Capacity utilisation models built from system logs informed hardware planning and budgeting.
Econometric model and contract valuation calculator under different economic scenarios supported the estimation of the value of key customer businesses. Provided strategic inputs and quantified required capital allocation for entity separation decisions.
Data science mentorship across NLP, computer vision, econometrics, and machine learning — with completed models integrated into business processes. Value-at-risk calculator enabled effective market risk management for an online trading platform.
Whether you're looking to start a new modelling project, modernise existing processes, or simply explore what's possible — we'd be glad to hear from you.