Today's internet is ruled by algorithms. These mathematical creations determine what you see in your Facebook feed, what movies Netflix recommends to you, and what ads you see in your Gmail. Engineers at companies like Facebook and Google spend huge amounts of time developing new algorithms---and tweaking old ones. Yet some academic computer science researchers spend years developing algorithms that are seldom seen outside of academia---even if they could be of great use to engineers in the private sector.
Diego Oppenheimer was all too aware of this as a program manager at Microsoft, where he helped design data analysis features for tools such as Excel and Power Pivot. He was always searching for better algorithms to integrate into these applications, and often found the answers he needed at Microsoft Research, the company's blue-sky research division. "I would find that people had been working on these algorithms for years, but we'd never heard of them," he says.
Meanwhile, Oppenheimer's college friend Kenny Daniel was working on his PhD in artificial intelligence at the University of Southern California. He had published multiple algorithms that were well received by academics, but had little chance of making their way into real-world applications. So the two of them teamed up to solve their mutual problem. Their answer is Algorithmia, which is essentially an "app store" for algorithms.
>'I would find that people had been working on these algorithms for years, but we'd never heard of them.'
The idea is to give algorithm creators the chance to have their work used in the real-world, and get paid for it, while making it easier for companies that don't have the resources of Microsoft or Google to tap into the world of algorithm development and find the best solutions to their problems.
There are a few other algorithm marketplaces out there already, including DataXu, which offers algorithms for ad placement, SnapAnalytics, which specializes in selling pre-made predictive models, and LumenData's Algorithms.io, which focuses on machine learning algorithms. But Algorithmia is different in that it will accept and sell any type of algorithm.
Creating an algorithm market is more complicated than it might sound. Oppenheimer explains that even if you find an algorithm you want to use in an academic paper, it's often difficult to actually implement it. Licensing the use of the algorithm to customers is a challenge as well.
Algorithmia is solving both problems by offering access to the algorithms through an online service. When you find an algorithm you want to use, you'll just add some simple code to your application that will route your query through Algorithmia's computer servers, eliminating the need to integrate the algorithm itself into your application. The cost for each of these queries will depend on the prices set by algorithm creators themselves.
The creators will also have the option of making their algorithms open source, so developers will be able to integrate them directly into their own applications---without a cloud service---but Algorithmia is hoping that the cloud service will provide enough convenience to make it worth paying for. The downside is that customers would need to send the data they need crunched to Algorithmia's servers for processing.
Algorithmia will host the algorithms on multiple cloud service providers, such as Amazon and Rackspace, which will speed things up for customers who use those particular hosts. And, eventually, Algorithmia will offer a version that companies can host in their own data centers as well.
Analytics consultant and trainer Ajay Ohri has been hoping for something along these lines for years. "For some time now, I had been hoping for a place where new package or algorithm developers get at least a fraction of the money that iPad or iPhone application developers get," he wrote on his site Decision Stats in 2011. "It is hard work to think of new algols, and some of them can really be useful."
He's not yet sure whether Algorithmia is the solution he's been waiting for. "It's too early to comment on Algorithmia since it is in private beta," he tells us. But he hopes it can grow into something like a "Yelp" for algorithms, which he thinks will ultimately be more important than the marketplace component itself because it's really hard for developers to determine which algorithms are the best to use for different applications. Thanks to its ranking and commenting features, Algorithmia may well fill this role.
