Now, an Indian search engine powered by supercomputer

Saturday, 22 March 2014 - 7:45am IST | Agency: dna

Scouring the net for a pink pair of shoes to match that killer dress you bought over the weekend. Well, it may help to think beyond Google and Bing., a search engine backed by a supercomputer and powerful algorithm, promises to be a revolutionary image search website that can rummage through online stacks to find that desired product.

The high-end search engine has been created locally by an Indian firm, which claims it can beat the global players.

The website has over 200 million products currently available from over 10,000 merchants.

Almost five years in the making, this is the fifth iteration of the product. Fully bootstrapped, self funded, operated in stealth mode for five years.
Next, the search engine plans to offer highly advanced travel and local search. offers a deep learning e-commerce search engine with a unique interactive User Interface and relational natural language search capabilities promising "any which way search", said Sukhbir Benipal, CEO of

It does by understanding the user query and refining search results to show only the most relevant products.

This search is very different from keyword search being offered by most search engine providers today, he said.

"I assembled a 10 TFlops distributed cluster with 20 nodes, 640 cores, 7.6 TB RAM, 10G networking, 25 TB SSD and roughly 1 PB HDD storage, that is already straining hard under current load. The current set-up can be expanded to 1 PFlops with additional CPU and GPU processing power. That would make it a pretty huge supercomputer, if I went into parallel processing with Rocks on it," Benipal said.

He claimed his search engine is "much more advanced and highly evolved (when) compared with IBM's supercomputer Watson and other cognitive systems and search engines. It runs on self built hardware at tenth the cost of Watson. (It is) tested to beat both Google Shopping and Watson in any type of search, any type of query."

Although the search's success is based on Big Data. "In no way is any personal information about a user collected" Sukhbir claims. "All user data is totally stripped of any identifying parameters before being processed. In my set-up I have no use for anyone's personal information anyway." he said.

On the future course, he said, "When I started five years ago, I had no idea I would be where I am today. So I am not going to venture any guesses for myself. But it will definitely be in cutting edge technology competing with the best. As for, it will be competing with Alibaba, TaoBao and eBay, globally."

Jump to comments

Recommended Content