3 Takeaways from Sameer Maskey on AI at Work

Posted by Alyssa Verzino on Feb 1, 2019 10:16:42 AM

lightbulb

From chatbots invigorating customer service, to self-driving taxis, to medicine-delivering drones, the applications of AI are dazzling. Behind each company pushing forward a brilliant AI innovation, there stands a strong data science team. That’s exactly what Fusemachines focuses on: offering AI engineers to companies interested in building AI systems.

On Episode 25 of AI at Work, we sit down to talk with Sameer Maskey, CEO and founder at Fusemachines. He shares insights on recruiting AI talent, how AI is helping financial institutions, advice for executives on launching their first AI projects, and where the future opportunities for AI applications lie.

How to Hire AI Talent

Fusemachines certainly has extensive expertise in finding talent. When companies want to find and build an AI team in order to add new features to their product line or build an AI product themselves, Fusemachines finds talent, trains them, and builds a team.

There’s a unique social mission embedded in Maskey’s work. He recruits talent from underserved communities in the developing world and the United States, trains them in the skill sets critical for AI and connects them with opportunities in the US and Canada.

Hiring AI talent is no easy feat - the demand for machine learning engineers vastly outweighs the supply. Maskey shares that when identifying talent to train, he looks for engineers who are very strong in math.

He explains his reasoning, “a lot of the off-the-shelf tools also work, but they need to be heavily tuned. In order to really tune it though, you need to understand how the algorithms work. In order to understand how the algorithms work, you need to understand the math behind it. So I usually look for engineers who are quite proficient in math. And, obviously you need to be able to code.”

AI, Chatbots, and The Evolution of Customer Service

Recently, Maskey wrote an article in Forbes about how AI can help some of the problems facing financial institutions, covering chatbots, personalized customer service, fraud detection, process automation, and other topics.

In particular, he sees the domain of customer service as particularly receptive for chatbots, given that there is such an information overload. “A lot of people do want to converse with customer service reps, and be able to ask questions and get answers instead of having to do searches or wait on the customer service line for a half hour, Maskey elaborates. “I think there is quite a bit of potential for using chatbots to enable customers, the end users, to access information quite quickly.”

At the same time, however, he notes that “language is one of the hardest problems in machine learning.” Longer text is easier for machines to understand, while short sentences are harder. When building a chatbot, you have to be careful about how well it will work for your domain.

AI Advice for Executives

An executive hoping to see a particular outcome from an AI project should be aware of the importance of data. Maskey’s first piece of advice is to know all of the possible data you have access to, and to set up an infrastructure that enables the team to get the data they need. “Without the right kind of data, you won't be able to train good machine learning models, and you won't get the expected output from the machine learning system,” he explains.

His second point is to set appropriate expectations. Unlike typical software engineering projects, machine learning endeavors typically don’t fit in neatly into a predefined timeline. Rather than approaching this with a “we should have our first demo in six months” mentality, it is better to embrace an “it may take longer, it may take shorter” approach, given the iterative process of refining a model’s accuracy.

***

Data is certainly king. The success of machine learning engineers and data science teams depends on harnessing it, and exciting AI innovations are built upon on it.

Looking ahead, Maskey sees an untapped market for AI in developing countries. One exciting application he shares that’s already being utilized - delivering medicine to remote, inaccessible areas via drones. His predictions for AI in 2019 include the growth and proliferation of AI companies, as well as self-driving taxis. One thing is certain: we are on the cusp of many great transformations to come.

To learn more, tune in to this episode of AI at Work And, be sure to subscribe on iTunes or Google Play and share with your friends. 

New call-to-action

Topics: AI at Work