Since its early days, Rangle has always aimed to be actively engaged with the developer community both globally and locally here in Toronto, a city with the fastest growing tech-job market today. Having established the company as a leader in front end software development, we’ve recently been working on expanding into other areas such as machine learning. Now with more local resources to draw on than ever before, we want to work with the community to help clients around the world to realize Toronto’s excellence in this area.
Last Thursday Rangle was proud to host Toronto AI, one of Toronto’s largest AI meetups (with well over 3000 members). Dave MacDonald, one of the Toronto AI co-organizers, gave an excellent presentation on Keras, a high-level deep learning library that runs on top of lower level frameworks (such as TensorFlow) and makes it incredibly easy to get started with machine learning. After Dave’s talk, Xiyang Chen, a member of Rangle’s Machine Learning team, provided an overview of popular approaches to object recognition, focusing on the R-CNN and YOLO families of approaches.
It was a fantastic event. Both presenters did an excellent job of making sure that their talks were suited for people who are only familiar with the basics of machine learning, while also offering value to those who are more advanced. The audience was very engaged, which resulted in great discussion. With 100+ people on the waitlist this time, I am hoping we’ll have a chance to host even more people here at the next meetup.
Meanwhile, I wanted to say a few words about Rangle’s approach to machine learning and our plans for engaging with the Toronto AI community.
Rangle and AI
The most obvious avenue has involved delivering smoother UX via the single-page application approach. We’re replacing old school click-and-wait user interfaces reminiscent of the 1990s with web applications that are increasingly indistinguishable from desktop and native mobile applications.
Another avenue to accelerate product innovation is by moving logic to the front end and adopting architectures and technical practices that enable incremental changes to a product. (One of the latest iterations here is building full design systems.)
Rangle has played a key role in helping many organizations with this change and will continue doing so. This transition is far from over and a lot of work remains to be done. In particular, companies that have successfully migrated to the new front end stack and have adopted Agile will now need to learn how to innovate even faster, to keep up with their increasingly nimble competition. Modern DevOps and DesignOps will be increasingly important for that reason.
As we look to the future, however, we expect some of the greatest opportunities for innovation to increasingly come from a new direction: machine learning and artificial intelligence. In the last few years, we have seen rapid innovation in this domain, with new ideas being seemingly being published on a daily basis.
More importantly, there has been a strong trend towards those technologies becoming democratized. Even as leading technology companies engage in a talent war over the world’s top AI experts, it’s easier than ever to get started. Back when I was studying artificial intelligence at university, I remember working for weeks on a machine learning project to recognize handwritten digits, implementing individual nodes in C++, only to achieve moderate results. Today, as Dave showed at the above mentioned meetup, one can get decent results with just a few lines of Keras code, which can then be deployed on the cloud.
Those technologies will increasingly allow organizations to build products that are not only beautiful and attuned to customers needs, but also smart, liberating users from the tedious tasks and delighting them by anticipating their choices. We already see some of those opportunities demonstrated by companies such as Google, with its smart reply function, which predicts likely replies to an email, allowing the user to respond with just a single tap.
In addition to making existing products smarter, we can expect machine learning to bring about an entirely new set of products - things that weren’t possible or practical without machine learning, in much the same way that services such as Airbnb weren’t possible without the Internet. It may also require a rethinking of a lot of conventional products that would need to work in the world of AI. For example, what would gas pumps look like in the age of self-driving cars?
Engaging with the Toronto AI Community
As Rangle has been building our machine learning practice, we’ve felt blessed to be doing this in Toronto. This city is increasingly recognized around the world as one of the key global centers of machine learning. This can be attributed to both University of Toronto’s traditional role in academic research in this area (which has included some of the most important work ever done in AI) and a growing array of companies focused on commercializing machine learning. Toronto is really an exciting place to be for this work.
From the very beginning of Rangle’s history, as we worked to establish the company as a leader in front end development, we strived to do so through active involvement of the community, both local and outside Toronto. We hosted meetups (our longest running one turned 5 years old in June!), sponsored conferences, organized and attended events. We’ve always looked to do so in such a way that we would both learn from the community and share what we know ourselves.
This approach makes even more sense as we build our machine learning practice, which gives us the opportunity to leverage local resources more than we ever could before.
Please reach out to us with any ideas of what Rangle can do to support machine learning in Toronto, learn from the community and share our own insights.