Did you know that even data scientists can get frustrated by the complexity of creating and deploying artificial intelligence and machine learning? AI and ML have a reputation for being a black box, which is a problem for businesses looking at implementing AI, for software companies creating solutions with AI and ML, and for those trying to explain how it works and benefits their organizations.
Kwame Asiedu has run into this problem throughout his career, as have his co-founders Amit Varma and Padma Polash Paul. Why? Because data science was new, and how to implement it required learning at all levels. This is why Kwame and his co-founders started Braintoy. They wanted to address the issues that they were experiencing as data scientists. ML is being built and integrated into most software solutions as we move into the future, but it isn’t the easiest thing to incorporate, explain or roll out.
Kwame started working as a business analyst and built his career on data analysis and modelling. With a background in computer science and geomatic engineering, he had the opportunity to work at Alberta institutions such as Enmax, Shaw and ATB with a focus on data science. In his early days, Kwame was often the only data scientist on board and building concepts on how to incorporate machine learning to the benefit of the organization. In each role, identifying what the data could do to improve operations was just one step. Educating the team he was working with, developing the team’s leadership, and ultimately, identifying how systems had to be built to support the effort were the next steps.
This is the pain point that Kwame and his team are looking to address.
Data scientists—especially those in small organizations, or small teams—have to identify not only the problem that is going to be solved for an organization, but how to gather the data and implement a model that works. This requires its own set of infrastructure and controls to ensure the project moves to the next stage.
Data scientists have to ask themselves:
How do they get the data and maintain/update it?
How do they build their model?
How do they check their model for compliance?
How do they roll out that model? Or make updates to the model?
In many companies, the infrastructure to support this is built from scratch, which makes ML expensive and extends the time frame for incorporating it. Kwame and his team experienced this in the industry and this is where Braintoy comes in. Rather than a company having to build the infrastructure itself, they can make use of their platform so that they don’t have to recreate the process.
An example of how Braintoy has been able to help a customer is with Calgary-based BoxofDocs. BoxofDocs is a search portal that offers municipal governments an easier way to search for documents, such as policies, requests for proposals and bylaws. But the company needed a way for its clients to search for documents based on specific parameters. Braintoy took an initial 150,000 documents and built specific parameters using ML to allow for improved searchability and allowed for the ongoing upload of over 10,000 additional documents a day. Braintoy’s platform provides a much easier way to search and classify than using Google.
Braintoy is working towards expanding the ease of use of AI and making it available to anyone who has data and a problem they need to solve. An easy user interface and the infrastructure in place to support clients as they move forward, are the first steps.
Kwame, Amit and Padma are all newcomers to Canada and each has been given the opportunity to move to other parts of North America. Data scientists with experience are in high demand so they could have relocated to Silicon Valley and gone to work for Google. But Calgary is home: it’s where they started their lives and they want to help their community by building their business here. Their goal moving forward is to make AI more accessible and less expensive and they want to start with Alberta businesses.
Check out our full interview: