Meet our AI Head, NAVEEN HONEST RAJ
You’re now meeting the brain behind our cool AI work.
-
How did you end up working for Skcript?
Well I was searching for startups as I was very keen on joining one. I didn’t want to join some big corporate enterprise, because I felt that being a part of a startup would help give me a better idea of the product flow. I started looking for places with job openings coinciding with my areas of interest and I came across Skcript. After the interview I knew that this was the place I wanted to work at. -
What is your favourite Google Product?
Tensorflow and Android. Tensorflow because all of my work is dependant on that basically. Android because I’ve been an Android enthusiast since the days of Honeycomb, and poking around the developer settings got me interested in Tech in the first place. Android has always fascinated me and it’s close to my heart. -
What made you interested in AI?
So I was basically was a backend developer. The thing that caught my attention in the world of AI was the Prisma application on the Google store. It used style transfer algorithms and I was intrigued by all of it so I started doing some research and I finally wound up in AI. -
What is the most innovative product you’ve seen using AI?
The Prisma app was really impressive and it still is. Recommendation systems are super cool as well. I think people haven’t understood how much AI has already advanced and how it’s implemented in quite a few places. Customer target ads have a lot of AI stuff going on behind the scenes. I think we’ll be seeing a lot more things with AI coming up in the future. -
What is the best thing you’ve learnt working at Skcript?
There’s nothing you can’t learn if you put enough time into it. I’ve learnt quite a few things here. I came here as a proper back end developer and I’ve managed to learn quite a lot about AI development. Before coming here I knew very little about AI and I’m really proud of how far I’ve come. -
Which software do you think is essential for what you do?
Tensorflow (without question). It’s so powerful at what it does, and it’s library is the most important part of it. All mathematical operations are expressed as small functions and what would otherwise take more than 80 lines of code, takes just 2. Without it, I wouldn’t be able to do most of what I do. -
Google Assistant or Siri? How do you see Google Assistant evolving in the future?
Google Assistant, hands down. I’ve been a fan of Android and Google Assistant just fits with everything. It was released recently, but has already grown to be very capable. Google is literally everywhere, on all our devices, and it’s become a part of all of our lives in one way or another. API control is easy and Google is just the way to go. Google Assistant is going to grow exponentially. -
What is the first thing you do as soon as you come to the office?
Well I turn on the Wi-Fi, and go through all the different types of social media. Then the fun begins. -
What motivates you everday to push forward? What motivates you to go to the gym?
Well as soon as I turn on the computer, there’s a painstakingly long task-list which greets me. There’s usually about 60-70% that needs working on and I just get started. I go to the gym to avoid chances of breaking up with my girlfriend 😅 -
Who at the office do you like the most? Why?
Dikson. We relate the most and we just understand each other. We’ve really bonded after coming here and he’s really nice to work with. -
How was your Tensorflow talk? Share your experience.
My learning experience with Tensorflow was great, and I just wanted to share my experience. I’ve spoken twice so far, once here at Skcript for GDG and another time at a college in Madurai. I was really passionate about giving a speech at college as I could really relate to them. They all knew about AI and some of what it can do, but they didn’t know about Tensorflow and how easy it actually is to make things. I told them how it’s very easy to implement it and make great things possible. -
If you’ve made something with AI, how do you measure it’s success? What do you benchmark it against?
I don’t have a survey or a model or something to compare it with, because what I make is usually highly specific. I always have expectations to what the software should be able to do, and once it reaches that level, I’ll be content. If I’m trying to implement NLP and it gives me the right output for what I think is acceptable input, then it’s a success for me. -
And lastly, are you as ‘honest’ as your middle name?
Well, I try to be (sometimes!) 😂