Predictive Modeling Engine:

We are currently working on products that will enable the clients to  understand the customer behavior /buying pattern in a typical retail scenario. We have incorporated latest technologies to make sense of this valuable data from analyzing and mathematically modeling the purchasing pattern of each customer. This product once developed will be of immense help to the retail stores enabling them to improve their sales significantly. Our modeling engine will be quantifying the qualitative data obtained from the customer. By forming a proper structure and a model to capture this data we will be feeding in this input to our modeling engine to predict, forecast the sales and the client can simulate various strategies to increase their revenue and see the output. The engine has a self learning algorithm, thus the accuracy improves with time as more data is being fed into it. Every Miniscule detail is carefully analyzed and given appropriate weight age, so that the data fed into our modeling engine is accurate. This helps us to get the output with great degree of accuracy.

"Intelligent Machine trading system"

We're working on implementing a distributed machine trading system, which unlike traditional models, relies on tagged news feeds, besides historic data. To reduce the hardware costs, the system is being developed to run on a network of consumer grade PCs, instead of big iron servers. Adding newer computers to the cluster automatically results on a improved data processing times. Further, in addition to a stochastic prediction system, we also implement a rule based system, which takes the inferences made by the predictor and filters them, based on user generated rules. All of this functionality is wrapped with a easy to use Web 2.0 style browser based interface, which runs on inbuilt web server in the master node.

Upcoming technologies:

Fault tolerant computing with Erlang.
The Hadoop distributed computing and storage framework.
Distributed Document oriented DBs with CouchDB.