Data Tree Solutions is a boutique provide of custom high quality solutions to data, AI, and software problems. With decades of experience in all parts of tech we make the impossible possible.

Data is an essential pillar of your business, but without expert handling it can be unwieldy and even lead you astray.
At Data Tree we are experts at building and handling petabyte scale datasets.
Whether you need ground up construction or just looking to squeeze extra value out of your existing data we can make it happen.
AI has made amazing breakthroughs in the past years, revolutionizing industries and enhancing efficiencies in ways previously thought impossible.
Data Tree has the expertise to harness those breakthroughs for your business, leveraging cutting-edge AI to analyze data, streamline operations, and drive innovation.
We can guide you through the complexities of AI implementation, ensuring that your business stays ahead in a rapidly evolving digital landscape.
Software makes data and artificial intelligence useful. Tying capabilities together for the end user.
We have experience building software in business of all sizes. One off prototypes to mission critical code deployed to tens of thousands of nodes world wide.
Deep experience in a variety of languages from C to Java, Perl to Php, Haskell to Lisp, Matlab or R, we have the experience to handle your needs.
“Their work save half an hour of time on the each of the dozens of buyer reports we produce every quarter.”
The best software is worthless without the people that use and maintain it. At Data Tree we belive software needs to be human connected. That’s why all our software is developed at our offices in Brooklyn, NYC.
Software developments should be a conversation with a friend not a corporatized game of telephone.
This project shows how our proactive forward thinking combined with a deep knowledge of AI produced a tool to reduce the impact of failures.
The video ad ecosystem is an extremely complex space. With dozens of interacting systems there is an endless variety of ways to for systems to fail.
In this environment question isn’t if failures will occur it’s when and how you react to them.
Identified a lack of insight into a rapidly growing subset of advertising and proactively addressed it
This projected demonstrates our ability to develop novel solutions to complex problems. In addition a strong mathematical background (beyond that of a typical software engineer) allowed for accurate modeling of the solution prior to implementation.
An early obstacle in the adoption of video ads, was that publishers (e.g. NY Times) were only paid when the ad (e.g. a 30 second car ad) started playing. This left the advertiser (e.g. Ford) without a finical cost if the ad failed to play for any reason (e.g. Ford misconfigured their ad server). When combined with programatic real-time-bidding this setup the possibility of large amounts of lost revenue.
In a small system (i.e. 10s of nodes) this would be trivial to solve. When scaled to 10,000s of nodes, trivial solutions fall apart. We designed a novel system, that took inspiration for Internet congestion control algorithms, that allowed for a fully distributed global calculations to be done instionaniouslly.
The system was molded with differential equations to prove its effectiveness prior to implementation. Deployment was flawless and the system remained in -use largely unchanged to this day.
“The first thing I do everyday is look at dashboard they created!”
Aparna Pappu Director of Ad Exchange @ Google