LightTwist Software – Cutting Edge Scientific Software

LightTwist Software is the company of Dan Ababei, an experienced software developer who learned his trade while wandering through such landscapes as the schools of Bucharest, Romania, the commercial sector in the U.K., the Dutch and the Australian academic world.

The company was created in December 2008, in Delft, the Netherlands. It has migrated, at the end of 2014, to Melbourne, Australia.

We undertake projects in any software field, but confess to a strong bias towards scientific software, especially involving pioneering science.

Current projects include

• Cutting edge dependence modelling and risk analysis tools: dependence trees, regular vines and continuous non-parametric Bayesian networks. See Uninet in the software section for details.
• Creation and development of FROST, a wildfire regime simulation tool, and work on the underlying simulator Phoenix, for the University of Melbourne, Australia.
• Work on geological interpretation tools: spectral analysis of wire-line logs, with a focus on orbitally forced climatic change (the Milankovitch Theory). Under contract from Enres International, in the Netherlands. Implementing new features and modernising the codebase for more than five years, as well as training and coordinating the other developers. See CycloLog in the software section for details.
• Creation and development of BBN Sculptor a support tool dedicated to exploring a large number of conditionalisation profiles for Bayesian networks created in Uninet, under contract from the National Institute for Aerospace in Hampton, Virginia, USA.

Past projects include

• The mathematical software engine for CATS, a working causal air-safety model developed for the Dutch Ministry of Transportation. The model comprises the sub-system descriptions, data gathering and software necessary in order to gain insight into the working of the safety system of air transport and to develop improvement strategies.
• The mathematical software engine for BENERIS, a EU project whose objective is to forge major advancements in food benefit-risk analysis on human health. The project utilises a benefit-risk approach with an iterative top-down means to explore risks of food and its contaminants.
• Full implementation of the mathematical engine for a web-based application facilitating remote Structured Expert Judgement sessions. Based on the Cooke classical method, under contract from the Center for Disease Dynamics, Economics & Policy (CDDEP).
• Design and implementation, from the ground up, of a powerful risk modelling tool for oil processing plants; a model represents a tree whose leaves are Bayesian networks (tens of thousands of networks can exist in one model). The program can sample and apply conditioning profiles to such a model in real-time. Under contract from the Technical University of Delft, in collaboration with Shell.