Software Developer in Penn Valley, PA, United States

I designed and developed a genetic algorithm-based optimizer that can find global minima or maxima in an n-dimensional hyperspace. This was written in CUDA to parallelize and find a solution quickly.

I designed and developed from scratch a distributed disparate data architecture that users could access with a simple SQL interface (or API if they prefer). It’s a distributed architecture—with modules written in C++, CUDA, Perl, .NET, and Java—that is spread over all servers at Rosenblatt and seamlessly scales up in the cloud as the demand goes up.

The data was stored in RDBMS, NoSQL, and flat files. Users need not be aware of data location or schema.

Data Manager takes care of corporate actions and returns data either adjusted or otherwise as per the user’s request.

RSIGrid encapsulates real-time and historical market data, derived statistical data, corporate actions, portfolio analytics, trading data, logging, authentication & access control, application monitoring, data archival, reporting, and more.

At heart of it is a multithreaded and distributed (master-master) messaging broker that implements several advanced. features like message priorities, resource starvation protection, pre-fetch, fair queuing, and more. The broker was designed. and developed by me from scratch in C++. The broker also reports stats for monitoring in a web-based state of the art. web front end. The front end was written using Web Sockets and HTML5.