Entering what is for me a brave new world of embedded development.

I have worked in a team that delivered a new portfolio optimization tool to an important client in 6 months (a completely new codebase). I have helped design and implement a interoperability layer around the (C++) implementation that makes it easy to create bindings to the optimizer in other languages. I have created complete binding libraries for Python (2 + 3), .NET, and Matlab. I have also been involved in 'CI-lite' tasks to support the release of client deliverables. Improving the robustness of our CI process by introducing LLVM tooling such as UBSAN, ASAN, clang-tidy and clang-format.I am currently working on re-architecting an MLOC legacy codebase using the techniques from the portfolio optimizer project that have been shown to add value. This includes adding the interoperability layer with python bindings with the addition of code generation . We are also replacing Perl scripts with python equivalents.

I returned to Spiral to start working on the refinery planning and scheduling product line. I worked on a team fixing issues that were blocking the global rollout of these new tools to a global oil major.

I moved to a completely new problem domain – design of microchips and joined the clock tree synthesis group of one of the leading products in this area. I was involved in bug fixing and also implementing a new algorithm for use by a leading client that allows for lower power designs to be created.

I have setup (from scratch) and am leading a new quantitative development team in SunGard's Cambridge office. We are working on several greenfield projects but concentrating on producing performance attribution tools that can be shared across the enterprise.

I was the lead developer in a major overhaul of the crude modelling for a global oil major. This involved enhancing existing functionality (much of which I wrote) as well as adding completely new modelling tools. Prior to this I was the lead maths developer in a project to port all the crude modelling of the world’s largest oil major to Spiral’s toolset. Much of the client data came from proprietary laboratory methods and required a lot of data analysis to create adequate modelling. Other data was industry standard but with features that needed understanding before our existing modelling could be applied successfully.The implementation was in C++ with much of the automation and analysis in Python 2.7.

I managed a team of 5 doctoral level analysts producing models for risk analysis of exotic derivatives.

I implemented a VaR model that was used to produce daily risk reports across the entire London trading operation.I implemented a parametric yield curve model for the Gilt STRIPs market making business.I supported the day to day risk management/reporting duties on the trading floor.