It looks like you are in United States. Would you like to go to https://onsiter.com/us/ instead?
Peut être disponible
(Mis à jour 2020-06-15)Kvantitativ utvecklare
London,UK och Malmö, Sweden
Natif Svenska, English
- Finansiell risk
- C/C++
- C#/dotnet core
Compétences (58)
SQL
VISUAL C++
C#
C++
SOFTWARE DEVELOPMENT
BACK END
MS .NET
API
Risk Management
backend
STRUCTURED SOFTWARE
.NET
QUANTITATIVE
QUANTITATIVE ANALYSIS
data
.NET Core
C
SCRIPTING
CREDIT
Python
MONTE CARLO
STATISTICS
CONTRACTS
ETL
TRADING
XSLT
Excel
HTML
JavaScript
XML
VBA
PRICING
DELPHI
FRONT END
Perl
Scientist
MATLAB
FRONT-END
Pascal
ORDER MANAGEMENT
COLLECTION
OFFSHORE
PUBLISHING
TOOLING
Fortran
PRODUCT DEVELOPMENT
ACROBAT
ONBOARDING
ENGINES
WELDING
Java
PIPELINE
TRANSLATOR
QA
VENDOR SELECTION
ERP
MAINTENANCE
Résumé
For Polestar the work has been around helping a freshly formed team get off the ground. The technology stack is dotnet core on AWS for the back end, react for the front end.
The work at UBS from 2017 onwards has been focused on optimization of existing calculations, pre-trade risk, collateral optimization, and further integration of other libraries for the XVA quant team. Accessing pre-trade information for certain clients has been reduced from hours to minutes.
At ADIA, after setting the team up there were two streams of work; Short term projects that brought immediate value, e.g. mobile apps and gamification of fund management. The second stream was longer term and aimed to deliver a SOA architecture to bring in heterogenous data sources under a homogenous interface.
The work for UBS up to 2014 centered around the trade description language PAL for the XVA quant team, and closely related functionality, such as parts of the AMC and trade translators as well as library architecture and setting up regression test suites.
Expérience professionnelle
Emilix Ltd
2018-11 - Actuel
Product development for data product for Emilix Ltd, gathering data on illicit streaming and downloading of movies and TV programs. This is written in C++ with supporting functionality in python (maintenance scripts and an administrative interface written in Vue/JS/HTML, with a python backend based on Tornado). A Postgresql database is used as the data layer. Parts of the database is a traditional relational design, and a third-party database is imported into document style tables. This is all running on Ubuntu. The data gathering software is receiving data over UDP, from the low ten-thousands of peers. The database is expected to grow to 8 TB over 12 months. Using docker for deployment.
November 2019 -Ongoing Emilix Ltd -Polestar
Build AWS serverless backend based on .Net Core for Polestar, using DynamoDB and ElasticSearch for data storage and search capabilities. Setting up ElasticSearch, development in C# and python. Coordination with UI team and some stakeholder management. Have also set up an onboarding and training process for new team members, two people onboard in January and two people in May. Detailing testing strategies with QA and providing reference implementations. Expected end of project: 2020/06/30.
November 2019 -Ongoing Emilix Ltd -Polestar
Build AWS serverless backend based on .Net Core for Polestar, using DynamoDB and ElasticSearch for data storage and search capabilities. Setting up ElasticSearch, development in C# and python. Coordination with UI team and some stakeholder management. Have also set up an onboarding and training process for new team members, two people onboard in January and two people in May. Detailing testing strategies with QA and providing reference implementations. Expected end of project: 2020/06/30.
Emilix Ltd - UBS
2017-09 - 2019-09
Emilix Ltd - UBS
Again working with the XVA quant team, I have been working on performance issues that have cropped up over the last few years. In this work I investigated some opportunities for moving to CUDA, and how it can be called from various runtimes (.Net and python being of particular interest) and identified, rectified and implemented a set of changes reducing compute capacity requirements, in some cases by close to 40%. This work took place in C# and some of the heavier number crunching was moved to C++, optimized using icc and VTune. I wrote a system for pre-deal risk profiles of hedge funds, giving the credit officers tooling to get relevant information in a timely manner, from a counterparty/netting pool level down to individual trades. The front end was written in Vue/JS/HTML, the backend was based on .Net and MongoDB. Path-sharing between different risk engines has been set up for FX and equities. I have also written the infrastructure required to move to bank wide curves and an implementation of the old interface back with the new curves, to help with the move to MCOF. At the end of the contract I was involved in two streams of work
porting the library to .Net Core, and preparatory work to move functionality in under OneApi for use in Jupyter notebooks as well as plain python. Other ongoing work include exposing the batch oriented framework into python as an API, as well as ad-hoc scripts for various purposes, typically written in python.
Again working with the XVA quant team, I have been working on performance issues that have cropped up over the last few years. In this work I investigated some opportunities for moving to CUDA, and how it can be called from various runtimes (.Net and python being of particular interest) and identified, rectified and implemented a set of changes reducing compute capacity requirements, in some cases by close to 40%. This work took place in C# and some of the heavier number crunching was moved to C++, optimized using icc and VTune. I wrote a system for pre-deal risk profiles of hedge funds, giving the credit officers tooling to get relevant information in a timely manner, from a counterparty/netting pool level down to individual trades. The front end was written in Vue/JS/HTML, the backend was based on .Net and MongoDB. Path-sharing between different risk engines has been set up for FX and equities. I have also written the infrastructure required to move to bank wide curves and an implementation of the old interface back with the new curves, to help with the move to MCOF. At the end of the contract I was involved in two streams of work
porting the library to .Net Core, and preparatory work to move functionality in under OneApi for use in Jupyter notebooks as well as plain python. Other ongoing work include exposing the batch oriented framework into python as an API, as well as ad-hoc scripts for various purposes, typically written in python.
Abu Dhabi Investment Authority
2014-11 - 2017-08
Initially continue recruitment and manage initial projects to establish the inhouse software development function. This included two mobile projects, using the .Net stack and Angular, as well as handling vendor selection of an EDM system. The next major project was to develop the data fabric part of a SOA solution, which was still ongoing at the time I left ADIA. This is again .Net based, publishing data using the OData protocol. The data fabric supports quick onboarding of new data as the process has been streamlined. I delivered the first external data provider, Bloomberg Server API and BPipe, through OData. Other projects include a portfolio construction gamification project meant to demonstrate the intricacies of portfolio construction to new entrants to the industry. The backend was written in .Net and the front end in JavaScript and HTML5.
A Technical Guide
Main client UBS
2011-01 - 2014-09
The XVA quant team at UBS needed help finishing their portfolio analytics language (PAL), part of their Risk Analytics library (RA). I was brought in to do so, and followed the outline given in "Modelling, Pricing, and Hedging Counterparty Credit Exposure: A Technical Guide". The grammar is written in Antlr, with C# as the target.
As part of an effort to move the entire bank to the common analytics library (CAL), I was tasked with porting PAL to CAL, and with that support PAL under both Linux and Windows. At the end of that contract I was kept on to maintain and extend PAL, take care of trade translators, in particular Murex translators, and provide library architecture advice for RA and to generally keep an eye on code quality. This includes adding functionality to PAL to support various trades, improving and extending the grammar.
I also developed a language-to-language translator for the FX exotics trade description language, where I bridged the functionality of the FX analytics library. A second language-to-language translator for the Equities trade description language was developed, where I similarly have to work in the equities library to enable the translation. Various bug fixes, such as in the collection of observables for the AMC, deferred cash flows, and generating the correct observables for Asian options.
When structured trades have not had any suitable translators, I have handcrafted PAL based on the term sheets. The various pieces of work are done in C#, C++, PAL, python, and various DSLs. Typically foreign libraries are written in C++, so any work there has to work under Windows and Linux.
I also worked on the integration of Murex trades into the counterparty risk library. This work was centered around MxML and FpML, and setting up a suitable pipeline for getting the trades in RA.
As part of an effort to move the entire bank to the common analytics library (CAL), I was tasked with porting PAL to CAL, and with that support PAL under both Linux and Windows. At the end of that contract I was kept on to maintain and extend PAL, take care of trade translators, in particular Murex translators, and provide library architecture advice for RA and to generally keep an eye on code quality. This includes adding functionality to PAL to support various trades, improving and extending the grammar.
I also developed a language-to-language translator for the FX exotics trade description language, where I bridged the functionality of the FX analytics library. A second language-to-language translator for the Equities trade description language was developed, where I similarly have to work in the equities library to enable the translation. Various bug fixes, such as in the collection of observables for the AMC, deferred cash flows, and generating the correct observables for Asian options.
When structured trades have not had any suitable translators, I have handcrafted PAL based on the term sheets. The various pieces of work are done in C#, C++, PAL, python, and various DSLs. Typically foreign libraries are written in C++, so any work there has to work under Windows and Linux.
I also worked on the integration of Murex trades into the counterparty risk library. This work was centered around MxML and FpML, and setting up a suitable pipeline for getting the trades in RA.
Bank Vontobel's
2009-06 - 2010-10
investment bank is market making in a wide range of equity derivatives. Vontobel has an in-house trading system T01, which is written in C++ and python, to handle execution and order management on various exchanges, which forwards these trades to Front for risk management purposes. In T01 I worked on FIX, TCP and API-based connectivity to new trading places, such as CATS-OS, CATS-LS, and Bloomberg, and associated order and trade flows. For Front I implemented various risk reports as specified by risk management. I also developed a price and risk feed for their web based platform, derinet.ch.
Standard Chartered
2008-12 - 2009-06
The work covered trade capture (FABS) and risk reporting. Instruments and trades were defined in Excel spread sheets; most of my interaction with actual Excel this was to figure out why a particular sheet did not price and calculate risk as expected, typically because some input was missing or invalid, or that the trades were not exported or updated properly in the Sybase database for the overnight risk. I implemented a set of risk reports in C++ running on a DataSynapse grid, like pin risk on barriers. Mostly I took care of a set of problems that caused various existing reports to throw unexpected exceptions triggering re-runs, and quite a lot of refactoring to make maintenance easier and to reduce build times which was an issue.
KBC Financial Products
2007-06 - 2008-12
C++, Python and Oracle on Windows and Solaris. Some C#. I started out in KBC AIM, where I worked on Front. Most of the work I did on Front was moving in-house models into Front and risk reporting (mainly for HVAR) for a wide variety of instrument types, with a slight slant towards credit derivatives. I was then transferred into the high performance computing team to work on a grid abstraction layer (GAL) for grid computing (which supported Platform Symphony, the in-house Tibco RV-based grid or an in-process grid that was simply a representation of a threaded approach). After that I worked on adapting one of the Monte Carlo based equity derivative models for the GAL. The initial release of the GAL had interfaces to Excel, Front and Imagine. I was then tasked with moving the model into Sophis for the equity derivatives desk and adapting Sophis to use the model supplied Greeks. This required a separate editor application to be written (done in C#) to make entering them a viable proposal. Communication between the editor and Sophis was done using Tibco RV. After that I worked mostly on moving more in-house models into Sophis (own vanilla option models, variance swaps, PDE-based callable notes), implementing scenarios (rho bucketing) and adapting scenarios to report results to the firmwide HVAR.
Developer
Credit Suisse
2007-01 - 2007-05
in the equity exotics front office team. Excel, VBA, some C++ and C#. Maintaining sheets for various trading strategies and products.
Mako Global Derivatives
2006-01 - 2006-12
Mako is a market market maker on index options. I designed, coded and maintained a historical database and API for quantitative analysis. Other tasks included ongoing work on Mako's inhouse developed trading system, a highly specialized solution for market making written in C++
adviser on RDBMS-based solution
Bloomberg
2003-10 - 2005-12
Fixed Income Electronic Trading
Working on a live trading platform, enabling clients to provide pricing, handle RFQ's and orders.
Coordinating and leading front-end AutoEx development for the Bloomberg Swaptrader product. Four developers reported directly to me.
Other responsibilities include acting as adviser on RDBMS-based solution for own and other departments in the London office, member of the database technologies working group, reviewing and implementing protocol specifications and participating in technology reviews to provide feedback on solutions and how to efficiently develop and deliver them.
Working on a live trading platform, enabling clients to provide pricing, handle RFQ's and orders.
Coordinating and leading front-end AutoEx development for the Bloomberg Swaptrader product. Four developers reported directly to me.
Other responsibilities include acting as adviser on RDBMS-based solution for own and other departments in the London office, member of the database technologies working group, reviewing and implementing protocol specifications and participating in technology reviews to provide feedback on solutions and how to efficiently develop and deliver them.
Creative Media Group
2001-03 - 2003-10
Own business
Initially the work for about six months was a continuation of the work for Witchity, but on a consulting basis.
The major contract I worked on was FlexTP, a file sharing solution for Creative Media Group (CMG) which was worked on from early 2001 to the end of 2002. CMG sold off the product to Diino who now sells the solution.
The server side was written in C++, and cross-platform compability was carefully considered. It was a multi-threaded solution with strong encryption. It was tested on Linux, Solaris and MacOS X where Linux was the main platform. The client was written in Delphi, aside from the encryption routines and protocol engine, which were written in C++ and provided through a DLL.
The second most significant client was AB Mando, a company whose main product was a treatment for eating disorders such as anorexia and bulimia. I was brought in to help their lawyers to audit a solution for patient journals delivered by a 3rd party vendor, which had scalability and functionality problems. After they had obtained the source code, I helped fixing the issues and add new functionality.
Other shorter work was performed for KnowIT Consulting, training consultants in Sybase, AB Jerntorget, to deliver a proof-of-concept for XML to screen print quality PDF and PostScript (C++ and Perl) in 2002, finalisation of parallel publishing system for Elanders Solutions in 2001 (C++ and Interbase), order confirmation delivery module for IF LuftFilter AB in (Perl, MS SQL Server and HTML).
Initially the work for about six months was a continuation of the work for Witchity, but on a consulting basis.
The major contract I worked on was FlexTP, a file sharing solution for Creative Media Group (CMG) which was worked on from early 2001 to the end of 2002. CMG sold off the product to Diino who now sells the solution.
The server side was written in C++, and cross-platform compability was carefully considered. It was a multi-threaded solution with strong encryption. It was tested on Linux, Solaris and MacOS X where Linux was the main platform. The client was written in Delphi, aside from the encryption routines and protocol engine, which were written in C++ and provided through a DLL.
The second most significant client was AB Mando, a company whose main product was a treatment for eating disorders such as anorexia and bulimia. I was brought in to help their lawyers to audit a solution for patient journals delivered by a 3rd party vendor, which had scalability and functionality problems. After they had obtained the source code, I helped fixing the issues and add new functionality.
Other shorter work was performed for KnowIT Consulting, training consultants in Sybase, AB Jerntorget, to deliver a proof-of-concept for XML to screen print quality PDF and PostScript (C++ and Perl) in 2002, finalisation of parallel publishing system for Elanders Solutions in 2001 (C++ and Interbase), order confirmation delivery module for IF LuftFilter AB in (Perl, MS SQL Server and HTML).
Vice President of Technology
Witchity Capital Corporation
1999-01 - 2001-01
recruited and ran a development department consisting of four developers for Witchity Capital Corporation. The first product we developed, Online Call Manager (OCM), won an award for "Most Innovative Product Year 2000" at ISPCon, and another product, Klik2Talk, was awarded a patent, GB2363543.
The software was written in C++ using STL, and the MKS toolkit for porting NT environments, and ran on Solaris, Linux and NT.
The software was written in C++ using STL, and the MKS toolkit for porting NT environments, and ran on Solaris, Linux and NT.
Contractor
London for Diverse Interactive
1998-01 - 1999-01
Worked on two short contracts in the UK, one in London for Diverse Interactive working on their product ZoneWare, a toolkit for Delphi written in Delphi and Visual C++ for creating rich media solutions, and also for Subsea Offshore in Aberdeen where I made applications for managing their ROV's and associated data in their ERP system (Platinum) in Delphi, Interbase and MS SQL Server.
consultant and later senior systems architect
Linné Data
1997-01 - 1998-01
Skills used: C++, Visual C++, ATL, COM/DCOM, MS SQL Server.
consultant
Semcon Data
1997-01 - 1997-01
I was placed at a customer developing parallel publishing software for Ericsson Radio Systems. Skills used: C++, Delphi, Interbase, Oracle.
Parcours scolaire
Senior High School
Senior High School
2025-01 - 2025-01
Ebersteinska Skolan
Ebersteinska Skolan
2025-01 - 2025-01
BSc Hons
Open University
2009-01 - 2016-01
PgDip
University of Westminster
2020-06 - 2009-01
BSc
University of Skövde
1993-01 - 1996-01