SparkCognition: Detecting malware faster with machine learning

About SparkCognition

Based in Austin, Texas, SparkCognition is a global leader in cognitive computing analytics, offering a powerful artificial intelligence platform for malware detection and other applications.

Industries: Technology
Location: United States

SparkCognition uses Google Cloud Machine Learning Engine to increase the speed and scalability of its Android malware detection engine, DeepArmor.

Google Cloud Results

  • Detects new endpoint security threats and develops new features faster, improving customer satisfaction
  • Scales globally, instantly making new malware detection models available to customers
  • Continually improves quality by training data models on millions of samples

Detecting zero-day threats with 99.5% accuracy

Artificial intelligence (AI)—the ability for computers to augment human intelligence and solve problems quickly—has the potential to transform how we work and live. Recently, Google announced Google.ai, a division dedicated solely to AI advancements. Google Cloud Technology Partners, such as SparkCognition, are also pushing the envelope using AI to bring new capabilities to market.

SparkCognition applies its AI platform to large data streams, helping customers anticipate equipment failures, predict hedge fund performance, and detect new security threats. Its DeepArmor product is a machine-learning-based malware detection engine for Windows and Android that uses AI to help detect and prevent endpoint security threats at machine scale.

“With Google Cloud Platform, we can analyze new threats with sub-second response times worldwide. We reduced the training time for our deep learning and machine learning models, so we can adapt to new threats faster and stay ahead of the competition.”

Bryan Lares, Business Unit Lead, Director of Cognitive Endpoint Security, SparkCognition

SparkCognition chose to host its software on Google Cloud Platform largely because of Google’s ongoing commitment to AI and its development of the TensorFlow open source library for machine learning. SparkCognition joined the Google Cloud Technology Partner program and began using Google Cloud Machine Learning Engine (MLE) to train data models to detect new malware threats targeting Android based devices. Large-scale training on a managed cluster empowers SparkCognition to easily train quality machine learning models much faster, while Google Compute Engine helps the company get new customers up and running quickly with a cloud-based management console.

“With Google Cloud Platform, we can analyze new threats with sub-second response times worldwide. We reduced the training time for our deep learning and machine learning models, so we can adapt to new threats faster and stay ahead of the competition,” says Bryan Lares, Business Unit Lead, Director of Cognitive Endpoint Security at SparkCognition.

A powerful machine learning engine

Building AI-powered technology is faster and more convenient with Google Cloud MLE, helping SparkCognition accelerate development and deployment of its DeepArmor for Android solution. With a scalable and distributed infrastructure to train new data models, SparkCognition can make predictions about new threats at scale and provide the fast response times its customers demand.

For feature extraction in model training, SparkCognition uses Google Cloud Dataproc, a fully managed cloud service for running Apache Spark and Hadoop clusters, which allows SparkCognition to perform ETL and data vectorization quickly, easily, and cost effectively. TensorFlow provides a straightforward framework for SparkCognition to build and test deep learning models, while on-demand Graphics Processing Units (GPUs), capable of tens of teraflops of performance, accelerate model training. Seamless integration with Google Cloud Storage provides fast access to tens of millions of file samples from global threat intelligence partners, while batch prediction helps predict target values for potential new threats.

“Google Cloud Platform offers great tools for machine learning, helping us with everything, from data ingestion through processing and vectorization. For an AI-focused company like us, this sets Google apart from other cloud providers. We can cost-effectively train our Android malware detection models on millions of file samples,” says Bryan.

“The power of Google Cloud Machine Learning Engine is the ability to access capacity on demand and quickly scale to millions of users without large up-front investments or infrastructure management. We can focus on model development and prediction and develop new features faster.”

Bryan Lares, Business Unit Lead, Director of Cognitive Endpoint Security, SparkCognition

Fast global rollouts

Once the models are trained, SparkCognition uses Google Cloud MLE to immediately make them accessible to customers anywhere in the world—using global load balancing for optimal performance—without having to build, manage, and secure its own data centers.

“The power of Google Cloud Machine Learning Engine is the ability to access capacity on demand and quickly scale to millions of users without large up-front investments or infrastructure management. We can focus on model development and prediction and develop new features faster,” says Bryan.

Working smarter in the cloud

SparkCognition understands the power of AI, which led it to use G Suite tools, such as Gmail, Google Drive, Sheets, Docs, Calendar, and Hangouts, for communication and collaboration. G Suite uses machine learning to improve productivity, helping users complete forms faster, eliminate spam, and visualize data instantly.

“We use Google products wherever we can because they’re scalable, cost effective, and offer the kind of dynamic features you would expect from a company with such a strong foundation in machine learning,” says Bryan.

Partnering with an AI leader

As SparkCognition enhances its software to address emerging threats targeting smart appliances and other devices connected to the Internet of Things, Google Cloud MLE will accelerate development and time to market.

“Google has been an innovator in cloud since day one, and Google Cloud Platform gives us a lot of powerful tools. We’re in a strong position to add value to the AI work that Google has already done, and the Google Cloud MLE team has been a big help,” adds Tina Thibodeau, Vice President, Strategic Alliances & Channels at SparkCognition.

About SparkCognition

Based in Austin, Texas, SparkCognition is a global leader in cognitive computing analytics, offering a powerful artificial intelligence platform for malware detection and other applications.

Industries: Technology
Location: United States
Google Cloud Platform logo

12 Months FREE TRIAL

Try Kubernetes Engine, BigQuery, and other Cloud Platform products with $300 in free credit and 12 months.

TRY IT FREE
Google Cloud Platform logo

12 Months FREE TRIAL

Try Kubernetes Engine, BigQuery, and other Cloud Platform products with $300 in free credit and 12 months.

TRY IT FREE