VentureBeat + NVIDIA + GeekSys: Startup Shootout

Could one of these dozen startups be the next Oculus?

Redshift Rendering Technologies


This sponsored post is produced by NVIDIA.

A dozen rising stars in the tech world will soon take to the stage to compete for $100,000 and a shot at glory.

It worked for Oculus. Only a year after appearing at the Emerging Companies Summit, Facebook acquired the virtual reality headset maker for $2 billion.

Each year, we talk with more innovative startups around the world that use the power of GPUs to disrupt industries and grow their business.

That’s why we created the Emerging Companies Summit. It’s a showcase for their brilliant work. And it introduces them to an audience of investors and technology execs who can open new opportunities.

The number of applicants vying for the summit’s top prize tripled this year to more than 150 companies from 30 countries. We selected the best 12, which will compete onstage in the summit’s Early Stage Challenge on March 18 in Silicon Valley.

Each CEO will have four minutes to present, plus four minutes to take questions from a panel that includes: Savitha Srinivasan, partner at IBM Venture Capital; tech pundit Rob Enderle; Saeed Amidi, founder and CEO of Plug and Play Tech Center; and Jeff Herbst, vice president of Business Development at NVIDIA.

Moderating it all will be NBC Bay Area tech reporter Scott Budman.

When they finish, the panel and audience will vote for the winner, who gets $100,000 on the spot.

Those interested in attending the event will find details here, along with a link to register. Company descriptions and presentation times are here.

The 12 competitors in next month’s Early Stage Challenge are:



(Ireland) – Deep Learning, Game Development

Today’s video games and movies require big art teams to render and create artwork. Artomatix develops tools that can automate digital media creation, allowing a single human artist to do the work of a team.

Ersatz Labs 

(U.S.) – Deep Learning, Big Data Analytics

Ersatz Labs makes a GPU-based platform for machine learning and “deep learning.” Deep learning is a type of machine learning algorithm based on the branching neuronal structure of the brain. Ersatz makes this technology practical and accessible through a web-based user interface and API.


(Germany) – Computational Fluid Dynamics, Supercomputing

FluiDyna designs and develops integrated solutions with computational fluid dynamics software for a range of customers, including vehicle manufacturers and suppliers, process and chemical engineering industries, and pharmaceutical companies.FluidDyna


(Brazil) – Computer Vision, Big Data Analytics

This award-winning startup pioneered in-store retail analytics. It has large customers in Brazil, where it is developing a single, uniform platform.


(France) – Automotive, Video and Image Processing

Intempora develops advanced driver assistance and autonomous vehicle technologies middleware. Its modular platform supports high-bandwidth data throughput from multiple sensors. It currently works with transportation, robotics and automotive companies, including Airbus, Honda and Renault.

Insilico Medicine

(U.S.) – Life Sciences, Big Data Analytics

Insilico Medicine uses genomics and big data analytics for drug discovery, drug repurposing and personalized preventative medicine. Insilico Medicine utilizes NVIDIA GPU to solve aging and age-related diseases. It also provides services to pharmaceutical companies interested in drug repurposing, early project termination, companion diagnostic development, and improving clinical trials enrollment.

NE Scientific

(U.S.) – Medical Imaging, Visualization

NE Scientific (NES) is developing new technologies for medical image analysis, modeling, and visualization. NES is sponsored by the National Cancer Institute for developing novel surgical guidance platforms for cancer treatment, is involved in projects for grading prostate cancer with magnetic resonance imaging, and in assessing the lung function in the intensive care unit using low intensity electrical fields for sensing tissues.

Pythia Systems

(Israel) – Cloud Computing and HPC, Big Data Analytics

Pythia offers a high-performance, low-cost data warehouse for analytics and reports. Its pilot project with the Israeli National Bank performed 20X faster than the bank’s data warehouse system.

QM Scientific

(U.S.) – Machine Learning, Computer Vision

QM Scientific helps consumers make smart buying decisions in real time. Consumers tell its app what they need, and the company’s platform recommends a store optimized by price, quality, shopping behavior, and location.

Redshift Rendering Technologies

(U.S.) – Rendering, Media and Entertainment

Redshift is the world’s fastest, most flexible GPU-accelerated renderer. In use today at many visual effects studios, Redshift shows that GPU-accelerated rendering offers a cost-effective, high-performance alternative to CPU rendering solutions.Redshift Rendering3

Replica Labs

(U.S.) – Computer Vision, Video and Image Processing

Turn the phone in your pocket into a 3D scanner. Replica Labs uses computer vision algorithms to quickly turn video feeds from any smartphone into accurate 3D reconstructions.


(U.S.) – Big Data Analytics, Graph Analytics, Cloud Computing

SYSTAP builds highly scalable software solutions for big graphs. Its technology can process large graphs hundreds or even thousands of times faster than competing solutions, bringing the power of high performance computing to an organization’s most time-critical graph challenges.

The Emerging Companies Summit, held Wednesday, March 18, is part of the annual GPU Technology Conference, taking place at the San Jose Convention Center in Silicon Valley from March 17 – 20. The summit provides companies with an opportunity to present their technologies to potential investors, partners, customers, analysts, and media.


Posted on 02/23/2015

Brazilian retail software uses facial recognition to generate data on customers

Our abilities to collect customer insight in offline stores still lags behind all that e-commerce can offer, but a Brazilian start-up is garnering support for its facial-recognition product to identify demographics.


Brazilian digital startup GeekSys has developed facial recognition software that it says enables retailers to create qualitative and quantitative insights on their customers.

According to the GeekSys team the software, named AdBoxSense, can recognize gender, age, ethnicity as well as gauge mood through facial expressions at point of sale. 

CEO Luiz Vitor Martinez says he started the project with the aim of enabling retailers to better understand their audience.  

AdBoxSense is supported by USB, webcam or fully embedded cameras. A cloud-based, mobile or desktop compatible dashboard is used to manage and then visualize data.

Martinez affirms that privacy invasion is not an issue as the software doesn’t store video footage on clients - merely collects data that is not tied to images of people.

AdBoxSense is not only valid for retailers - Martinez cites other possible uses, such as airport check-ins. However, according to Martinez, the GeekSys team has faced great difficulties in convincing the market of the potential of AdBoxSense.

He says that, in general, the retailer does not know what to do with such information: "The culture of retail (in Brazil) is still not very analytical." 


Original source: link (

Author: Sam Cowie

Posted on 03/11/2014

Exclusive interview: ISTOÉ Independente!

It is our great pleasure to present to our friends, customers, suppliers and investors an exclusive interview granted to ISTOÉ Independente.


"To understand consumers, two young Brazilians created a digital system that recognizes gender, age and interprets the reactions of those who look at shop windows or buy goods on supermarkets. Check out how it works"


We register our thanks to the publicist Thiago Ermano (TPress Comunicação Integrada) for inviting and supporting us, to the journalist and interviewer Lucas Bessel (ISTOÉ Independente) for his professionalism and hospitality, to all the recording and video editing team and to the volunteers (Izadora, Mariana, Larrisa, Alan and Rodrigo) who have accepted to be analyzed and have their ages revealed. 

Of course, we thank Bruno Martins and Gustavo Maia who make the GeekSys's team special and to all the people who have been contributing to our growth and daily quest for excellence in everything we do.

Excerpt from the interview with Vitor Luiz (CEO) and Luccas Menezes (CTO), respectively responsible for business development and technology in the AdBoxSense project, at ISTOÉ Independente’s office.

The journalist and interviewer Luccas Bessel, at ISTOÉ Independente’s recording studio. 

During the interview we received a proposal to do a real test to demonstrate the quality of AdBoxSense results. The first volunteer was Izadora, who has had her age estimated with only one year error and gender correctly classified.

Mariana has also contributed to the test, with similar results.

Next, the journalist Larrisa had her gender and age successfully estimated without errors. Note that glasses do not impact on the quality of the results.

The journalist Alan has also helped and contributed with his nice smile to the test. Besides having his age estimated with a three-year error and his gender properly classified, it was another opportunity to verify the operation of the solution AdBoxSence on individuals wearing glasses.

Finally, the journalist Rodrigo participated in the test and has had his gender properly classified and his age estimated with a six-year error. Although six years is still an acceptable error, it is noteworthy that the information [acho que se colocar “results” fica melhor o sentido da frase] provided in this test are partial.

In practice, information like gender, age, mood and attractiveness are extracted from detected faces and  then submitted to a statistical process to refine the results (not applied to the tests performed at ISTOÉ Independente).

Oficial interview video link

Posted on 02/03/2013

Eureka 2012 at the CNT channel!

Check out the story aired on CNT channel featuring Bruno Martins’s interview presenting the AdBoxSense solution embedded on mannequins to assess consumer profile.

Official link

Posted on 01/08/2013

AdBoxSense at “Olhar Digital”

Check the report by “Olhar Digital” in which AdBoxSense is mentioned as a powerful tool to analyse consumers in real time, enabling Marketing to increase sales.

Check out the full story at:


Official video available at: link

Posted on 01/08/2013

Inspirational Quotes

Simple is always better

Make it better, not cheaper

You can do anything, but you can't do everything

We never fail, we just find ways that don't work

Let our insights change your mindset

Make people want to work for the company

© 2013 GeekSys Serviços de Informátida LTDA Back to Top