What Khiem says – an interview with a Machine Learning Engineer

Did you ever wonder what's in the mind of a Machine Learning Engineer? We asked Khiem a few questions to find out.

As a working student in Machine Learning for almost a year now, Khiem already made a major contribution to our projects and the team! We wanted to know from him how he came to his passion, Machine Learning (ML), his sources of inspiration, and what he would recommend to aspiring Machine Learning engineers.

Khiem, what inspired you to become a Machine Learning Engineer?

I have always been amazed by how Machine Learning has advanced in recent years, as well as how it is being applied in multiple applications to solve real-world problems. Therefore, becoming a Machine Learning Engineer allows me to catch up with the fast-changing pace of the field while bringing research ideas on paper into deployable and robust products in reality.

How do you keep up to date in this fast-paced and ever-changing field?

I think one of the most effective ways to keep yourself updated with the latest trends in the ML world is to subscribe to multiple news channels like newsletters or blogs relating to the fields. These platforms help you quickly and selectively catch up with all the news without having to scroll through thousands of publications every day. I also personally pay attention to large conferences in the field such as CVPR or NeurIPS and new projects from big AI labs in the world like Meta’s or DeepMind’s, which unveil the coolest state-of-the-art research. But at the end of the day, I think that it really matters to have hands-on experiences with these cutting-edge techs to have a really good idea of what is happening inside.  

What’s the most fascinating experience you had in your career? 

I think I will never forget the day I managed to deploy a Vision Transformer model to production at Sereact that runs in real-time and generalizes way better than any other models I have trained.

What’s your favorite AI tool you use in your daily life? 

Definitely ChatGPT. It helps me on plenty of occasions when I need an explanation of a complex concept or if I have a bug without any solution on the Internet. Even though it sometimes makes up an answer, I find that its output is pretty accurate most of the time, and it truly improves my work throughput.  

What subjects would you include in a one-day Machine Learning crash course and why?

I would recommend an introduction to the ML theory module, which gives an overview of basic ML algorithms, and another basic programming module with Python and a popular ML framework like Pytorch.

What advice would you give to someone who is just starting in the field of Machine Learning and wants to become a successful engineer?

I think it is important to understand the underlying theoretical aspect of the ML system you are trying to build. This does not necessarily mean that you have to understand every single line of math. It is more like the general big picture, how and why the system works, and how to modify it if you have a different input or need a different output. It is also important for one to grasp the basic knowledge of software engineering since the main goal of your job is to deliver the ML system to production, which in turn requires the engineer also to have the skills and mindset to develop a software product.

Thank you Khiem for this great insight into your mind! If you have any questions left, don’t hesitate to contact Khiem on LinkedIn.