What Kanishk says – an interview with a Senior Deep Learning Engineer

Starting as a working student at Sereact, Kanishk fascinated the team with his passion for maths. He decided to stay, write his master's thesis and become a full-time team member. Read more about his motivation and his secret to unwind in this interview.

What aspects of your job do you find most fulfilling or rewarding as a Deep Learning Engineer?

I develop self-supervised deep-learning models that don't require hand-engineered labels. This approach involves creating explainable math for optimization, which requires intensive thinking and brainstorming. It's rewarding to see the models develop new mathematical models and simplify existing ones during their optimization process.

Is there a particular project or achievement that stands out as a highlight of your career so far?

I recently submitted a paper to a very well-known scientific conference. My paper introduces a groundbreaking method for training deep-learning models that achieve exceptional generalization of object representations in images. This novel method reduces computation resources significantly more than the originally proposed framework, showcasing its outstanding efficiency. This achievement is a significant highlight in my career.

If you could have dinner with any person in the tech world, who would it be and why? What would be the one burning question you'd ask them?

Not anyone in particular. I would rather discuss with the professors who taught me theoretical math foundations for intelligent systems and learn how to think math better.

As a Deep Learning Engineer, you likely spend much time in front of a computer. So how do you unwind and disconnect from technology? Any favorite hobbies or activities that help you relax?

I paint with a Renaissance theme, which usually keeps me occupied as there are many fine details to cover. Furthermore, I like to barbecue and play football.

If you could possess any superpower, what would it be, and how would you use it?

I don’t need any as they come up with additional responsibilities (Quoting Uncle Ben from Spiderman)

How do you stay motivated and continuously grow in your profession?

Seeing my models and algorithms come to life in the real world through robotics is motivating. Having a background in robotics engineering, I enjoy the challenge of applying AI with robotics to solve practical tasks. These interdisciplinary challenges motivate me to learn outside work hours and stay updated with the latest methods. This continuous learning and application process has greatly contributed to my professional development.

What’s your favorite item in the Sereact office?

The beer crate and the bottle of Aperol in the fridge (it is summer).

Looking to the future, are there any specific areas within deep learning or AI that you are particularly excited about exploring or diving deeper into?

One Sunday, tired after playing football, I craved a bottle of chilled water from the fridge. However, my laziness prompted me to consider engineering robots equipped with AI that could assist me. Recently, I have been enthusiastic about applying large language models like PickGPT, which possess visual scene understanding integrated with robotics to tackle various tasks. This excitement has fueled my desire to delve deeper into this area, with the ultimate goal of developing Artificial General Intelligence (AGI) and creating a robot capable of fetching me a drink from the fridge.