I finished my MSc.
at Cambridge University Statistical Signal
processing Group where I was a part of the STELLA group
and was supervised by Arnaud Doucet,
where I learnt a lot about SMC and MCMC techniques from Arnaud. I was
also lucky to spend some time working with Pierre Del Moral in
Toulouse, where he gave me some valuable insights into theoretical
properties of SMC techniques. After this I moved to spend a fantastic
year as a research assistant in Univeristy of British Columbia where I
worked with Arnaud, Nando de
Freitas and Kevin
Leyton Brown as part of a joint set up between the statistics
department and the computer science department. I was located in the LCI. At the end of that year
I made a difficult decision to move back to Australia, where I began
working as a quantitative analyst in CBA. I work in Market Risk and
Operational Risk modelling, with particular focus on implementation of
a Monte Carlo engine for BASELL II requirements. I worked as a quant
for 1.5 years before deciding to return to my other passion research,
this was helped through the support of a scholarship from CSIRO.
I now work on a project with them 2 days a week related to commodities
modelling and the rest of my time is dedicated to my other passion in
research which is related to methodological development of Monte Carlo
sampling techniques.
Recently, mid 2006, I received an Australian Post Graduate Scholarship and so have begun a Ph.D. with Scott Sisson and Yanan Fan as co-supervisors and a second supervisor Ben Goldys at the University of NSW Statistics Department. I am also occassionally discussing with Robert Kohn in Economics department at UNSW on possible econometric applications and methodology. My focus for research is two fold 1) Trans-Dimensional sampling methodology development and application. This work will aim to involve methodology, theoretical and applied aspect of Monte Carlo simulations, in particular Sequential Monte Carlo and Markov Chain Monte Carlo in a general state space setting. 2) Commodities modelling, this work is jointly funded by CSIRO CMIS group and involves developing multi-factor models for commodities. In particular the focus will be on simulation techniques for estimation of spot prices using observed futures contract prices and other non-linear observation of options pay-offs. This takes one away form the standard regime of linear and Gaussian state space models with Kalman Filtering and Maximum likelihood and allows for competitive use of Sequential Monte Carlo approaches