Catching Up with Kevy’s New Chief Data Officer Felipe Castrillon on Artificial Intelligence
One of our favorite local startups Kevy recently brought on seasoned data scientist Felipe Castrillon as Chief Data Scientist. Felipe brings his expertise in Artificial Intelligence (AI) and Machine Learning (ML) to the Kevy email marketing platform – a new service called Kevy Analytics. Sitting on top of the Kevy platform, Kevy Analytics has been instrumental in driving revenue for early adopter eCommerce marketers by creating more personalized experiences for their customers.
Felipe shares what he’s up to with Kevy and how the Kevy team is taking eCommerce marketing to the next level.
How do you leverage AI/ML at Kevy?
At Kevy, we are just at the beginning of a big push towards AI. Talking to our business partners daily, we realized that marketing folks are having to do more and more in order to get the same results. There’s just too much noise out there to get a successful message across and marketing departments are being stretched thin. What we want to do at Kevy is empower these folks to help them reach their customers on a personal level and improve their brand loyalty. Our solutions, which include customer segmentation and purchase probability scores, help our partners focus on personalization while taking difficult tasks off their plate. We have seen open rates increase 2-4x, click through rates increase 2-5x and revenue 2-10x. And this is just the beginning.
What do you think is the future of AI/ML in your industry?
E-commerce is becoming more and more complex and the traditional generic email blasts will just land you in the spam folder. In a few years, I predict that personalization will be the only game in town as customers want to connect with your brand on a more personal level. That’s why we have positioned Kevy to be a leader when it comes to personalization.
What do you see as the biggest hurdle companies will have to get over as they are trying to leverage AI in their business?
Education and implementation are a very important barrier. Some companies are not even aware of AI. Others have no idea how to implement it into their organizations. As an example, most software engineering departments do not really know how to successfully manage or evaluate an AI project. There is a learning curve that must be overcome in order to move forward. Companies like Kevy can help to bridge that gap by de-mystifying AI so that companies can leverage it without large organizational changes.
Do you think that there are ways in which we are under-utilizing AI/ML?
I would not go as far as saying that there is under-utilization because technology adoption needs to move at a certain pace. AI models have matured very rapidly from an engineering perspective and huge tech companies like Apple and Google have successfully applied the technology to their consumer apps. However, small and medium-sized companies are still trying to figure out if there is any benefit to their bottom line and how to implement it from a technical and organizational perspective.
In which ways do you see AI affecting us the most over the next 2 years?
This is a nice follow-up to the last question as I believe that small businesses and medium sized businesses will become more open to the benefits of AI. I think we are going to start seeing a more widespread adoption of the technology in many industries that does not just include the guys at the top.
What AI/ML capabilities get you most excited?
I really love the tools that are available today as they are democratizing AI. The advances in cloud and GPUs plus the easy-to-use libraries have really lowered the barrier to entry. Whereas before, only experts with super-computers could run a large model, new tools are making AI more and more accessible to the masses. The only downside is people blindly applying models and trusting them without knowing their strengths and limitations.
What are some AI/ML resources that you use?
I follow the Towards Data Science blog on Medium. What I really like about this blog is the step-by-step on-hands code tutorials and the interesting discussions on AI. The best thing is that it provides free information, and anyone can start following since they have posts for all skill levels and interests.
What are some of the tools you could not work without?
For big data pipelines I am a big fan of Apache Spark. This helps us process large amounts of data quickly and efficiently. We use DigitalOcean Cloud which has been a great partner to scale our operations and pipelines. For modeling we use python-scikit for simpler and quicker models and TensorFlow for running deep learning models.
Do you think people are justified to feel scared of AI? why/why not?
AI is still in its infancy and its limits are unknown. As of now, these models are very dumb for most human-like tasks so I do not think the robots are going to take over just yet. But in the future, we shall see!
Thank you so much and good luck on your new role
Want to meet or catch up with Felipe?
You can reach out via Kevy’s contact page or on LinkedIn.