Computational alloy design has been used to develop a new maraging steel system with low cost, using Mn for austenite reversion and Heusler Fe2SiTi nm-scale precipitates to strengthen the martensite, avoiding high cost alloying elements such as Ni and Co. A pronounced ageing response was obtained, of over 100 HV, associated with the formation of 2–30 nm Fe2SiTi precipitates alongside the development of 10% Mn rich austenite, at the martensite boundaries with the Kurdjumov-Sachs orientation relationship. The precipitates took on different orientation relationships, depending on the size scale and ageing time, with fine precipitates possessing an <100>L21//<100>α orientation relationship, compared to larger precipitates with <110>L21//<100>α. Computational alloy design has been used for the development and demonstration of an alloy design concept having multiple constraints. Whilst in this case computational design lacked the fidelity to completely replace experimental optimisation, it identifies the importance of embedding Thermodynamic and kinetic modelling within each experimental iteration, and vice versa, training the model between experimental iterations. In this approach, the model would guide targeted experiments, the experimental results would then be taken into future modelling to greatly accelerate the rate of alloy development.