Interview with Luca Venturi

0
86


Luca Venturi is the author of Hands-On Vision and Behavior for Self-Driving Cars, we got the chance to sit down with him and find out more about his experience of writing with Packt

Q: How did you become an author for Packt? Tell us about your journey.

Luca: I have been scouted by Packt. They contacted with the proposal to write a book related to self-driving cars. I found it very intriguing, but when I checked the arguments that they wanted to touch, I got a bit scared, as I could not cover all of them. So we discussed having a second author, and I asked Krishtof Korda to help me. He was in my team when I took the “Self Driving Car Nanodegree”, and I knew he could have been the perfect person to compliment me and cover the topics that were too much outside of my competency. Luckily, he was happy to jump on this adventure!

Q: How long did it take you to write the book?

Luca: It took around 4-5 months of intensive work, during the evenings and the weekends, plus a couple of months for refining, but at a much lower intensity. Some chapters were ready in a week, but some took more than a month, mainly for the code.

Q: What kind of research did you do, and how long did you spend researching before beginning the book?

Luca: My research was mainly based on courses that I attended in the past, and on online resources. Before writing the book, I mainly verified that there was a simulator that I could use to create the datasets (as many of the public datasets are not free for commercial use). Then I was researching chapter by chapter. I was focusing on refreshing but also updating my knowledge. I can say I learned a lot while writing this book.

Q: What advice would you give to readers learning tech? Do you have any top tips?

Luca: I think the best advice is to study what you like and develop a career based on what you like. Tech is tough, and if you work on things that you like, not only you will most likely achieve better results, but you will also have a lot of fun in the process. The second advice, above all for juniors, is to never underestimate the extent that other developers go, to reach their goals. So, don’t give easily on a problem, because that’s how you reach the next level.

Q: Did you face any challenges during the writing process? How did you overcome them?

Luca: The real problem was the code, as integrating with the CARLA, the self-driving car simulator, has not always been straightforward. Here the solution is just “keep calm and try harder”. It took me two weeks to figure out why a neural network was so slow inside Carla. I also needed the neural networks to perform relatively well, and that hasn’t always been easy. If you have ever tried to “debug” the behavior of a neural network, you know what I mean. Sometimes they can perform decently (but not well) even with wrong data, and it can be tricky to spot the problem. While several times I could have made the network more precise just increasing the size of the dataset, I tried to achieve good results with a limited amount of samples, because in real life it’s not always easy to add samples.

Q. How do you keep up-to-date on your tech?

Luca: I have a variety of sources. To get a complete description of a topic, I think books are the best. Then I follow some RSS feeds (yes!), I read the news and I take online courses. The discussions with ex-colleagues, very senior developers, can help a lot to understand what can really work. For Deep Learning and Big Data Google, white papers can be very valuable. And when you are stuck or trying to assess technology, StackOverflow is invaluable. It goes without saying that Google is your best friend. Let’s not forget the official channels: every day I check what’s coming to the next OpenJDK, because it is never too early to plan on adopting a new technology.

Q. Do you have a blog that readers can follow?

Luca: I haven’t blogged much so far, but I plan to do it more in the future. Please feel free to follow me here: https://medium.com/@l.venturi

Q. Can you share any blogs, websites, and forums to help readers gain a holistic view of the tech they are learning?

Luca: For Deep Learning, Towards Data Science is doing really well. For Java, I am a fan of DZone, where I also wrote an article.

Q. How would you describe your author journey with Packt? Would you recommend Packt to aspiring authors?

Luca: They have been very supportive, even when things seemed more difficult than expected. They were very responsive, which makes things much easier. I am surprised by how smoothly things went, considering that we were, working in 3 time zones with large differences, completely remotely with two new authors, and during this crazy pandemic. Plus, I am not even a native English speaker! I can totally recommend working for Packt.

You can find Luca’s book on Amazon by following this link: Please click here

Hands-On Vision and Behavior for Self-Driving Cars – Available on Amazon.com