Jay Walters

What's Here

I've started yet another little project, a C# implementation of ONC/RPC. It's too big to just host here, so it's up on sourceforge as project remotetea.net.

I have started a blog, so I have someplace to leave details but keep this page for big news. You can get to it by following the link from the masthead. I am going to try and move the news bits into the blog, though as of now I'm not quite diligent enough keeping the blog up to date.

My interests are in the technology of portfolio optimization and asset allocation. This all started out with optimization, but I've since moved on to looking at technology on top or next to the optimizer as there are plenty of people who have built really good optimizers. Mine are still hanging around the open source project and are servicable, but probably on par with a simple graduate school project.

I have written a paper on the Black-Litterman model, deconstructing many of the usual papers and working examples with their data in the hopes that it will help other implementors of the model. On the asset allocation front, I've done similar things in terms of playing around with various optimization techniques to learn the math behind the process. I've also played with more exotic models like an Entropy objective function, Mean-Variance with Entropy constraint, Simulated Annealing with a cardinality constraint and few other odd bits and pieces. Another optimization method, based on the ordering, but not the exact values is interesting. I came across this in Satchell and Scowcroft's book, "Advances in Portfolio Construction and Implementation." There is a partial implementation in the code, but I have not yet been able to figure out how to visualize the results in risk/return space.

Finance Focus

Having landed a job in the financial industry several years ago, I became quite interested in the mathematics of the job. After three years of hard work, I earned my CFA Charter and am now Jay Walters,CFA. It was quite the learning process going through the program, and very satisfying to achieve the ultimate goal of receiving my CFA Charter.

The sourcecode for my "toy" optimizer is up on sourceforge and if you navigate down into the Finance pages you can run it as an Applet from my website.  In the end portoflio optimization is about the data, not the math, but it's been a journey to learn about non-linear optimization and I'd like to share what I found. I'm hoping to pull together a set of links to the best papers and maybe write something up myself to go along with the code.  We all have our dreams!

Database Design

I've been collecting some new material for my guide to developing database applications.  The activity of writing this document and attempting to capture some of my experience over the last many years was motivated by a recent article I read in IEEE Software addressing the issue that much of Software Engineering literature is dedicated to process, but not to the topic of writing software itself. There are of course books that regurgitate the various APIs and show you how to use them in the small, but how does one learn how to write an application that efficiently uses an RDBMS? Another interesting paper in IEEE Software spoke of evaluating third party software based on the quality of the data model. I certainly have some comments there!  See my paper to see what I have to say.

I've also become a little more interested in the more abstract issue of Data Management or Data Architecture in the Financial Services industry.  The constantcy of data, new data pouring in, old data can you get rid of it, and in the meantime getting the right data to the right application is something I have not encountered in my previous positions.