Just five months after announcing itself to Australia at AWS Summit Sydney 2019, SaaS startup DataRock is gearing up for its global launch - and the story of how it got there is a fine case study of how channel players can transform into vendors.
DataRock is a joint venture between Melbourne-based AWS partner DiUS and data science consultancy Solve Geosolutions. The company’s product is a SaaS solution based on AWS machine learning capabilities .
The concept of spinning off DataRock into its own business came about 18 months ago after AWS introduced DiUS to Solve. DiUS had worked with a handful of clients in the mining sector to build machine learning models.
DiUS found that while it was working with multiple customers, the work had many common elements.
DiUS' client engagement principal Frank Losinno saw the potential to bring DiUS and Solve together, spin out their respective IP and expertise and create a startup to sell licenses for the platform as an ISV.
“We were getting a lot of this work and there wasn’t anyone in the market globally doing this at the moment,” said Losinno. He added that DiUS was already looking to spin out more of its products, having recently developed solutions for the utility and energy sectors.
Since its soft launch earlier this year, DataRock’s customers uncovered another use for the platform: capturing and analysing the backlog of imagery and video that customers have left unprocessed for up to 15 years.
“We’re trying to be an agnostic imagery platform,” said Losinno. “So we’re doing photo and video at the moment, any type of integrated imagery that we can derive and detect something in that image to make a process or workflow better.”
Fast-forward to the present and DataRock is just six months out from its global launch at the PDAC mineral exploration conference to be held in Canada in March 2020.
The platform is based on AWS’s machine learning services including Sagemaker and Ground Truth using a serverless model. AWS Step Functions is used to orchestrate processing, while files and metadata are stored in S3 and RDS, enabling image analysis as-a-service. The platform also uses Google Cloud’s OCR service.