#Wheredidyoustart №26: Xuefei Wen

I stumbled into the digital world from university, somewhat accidentally. My dream was to be a teacher so I applied to study Biology at a Chinese university that had a solid reputation of cultivating teachers.

Unfortunately, I didn’t get the grades for biology, but I did for computer science. At that time my knowledge of computers was limited. I knew about online chatting and computer games, but not much else. It was a tough start to university as I had no idea what our lecturers were talking about (programming…?). But after writing and successfully running my very first program in C++ (a simple print 'hello world' within a for loop) I started getting into it.

When we learned about data structures, ACM algorithms, operational system, and all the maths, I was hooked. I enjoyed my first AI and machine learning classes so much that I now work in this area and can see myself still enjoying it in 30 years!
Just because something goes differently to how you expect it to, it doesn't mean that it's not going to be amazing and the best thing for you.

A good university course will equip you with the systematic knowledge and mathematical foundations you need for further development in different areas of computer science.

I would stick to the same route as it has given me a deeper understanding of AI, but I think learning C++ as my first programming language focused too much on programming details rather than algorithm design. I wish I’d known about open course sources like Coursera earlier, as it would have helped me find other explanations when I didn't understand things from lecturers or books.

Having said that, it's not the only route. I've also seen people with a maths degree work in computer science, and those without a degree become great programmers. There are so many learning resources and if you have the enthusiasm to learn you can make digital a successful career.

I wish I’d known about open course sources like Coursera earlier, as it would have helped me find other explanations when I didn't understand things from lecturers or books

I am currently an Artificial Intelligence Engineer at Crisp who use advanced tech and human risk analysts to protect brands, platforms, advertisers and kids from risks posted on social media. These can be anything from swearing and sexual exploitation to cybersecurity issues, influencer scandals and PR crises.

My task is to build automatic AI techniques that accurately and quickly detect those risks in images and videos.

I love my job and am passionate about machine learning, researching it, and keeping up with the latest advances in this area. I’m always looking for new techniques that excite me and I get plenty of opportunities at Crisp to research resources, attend conferences and learn from the excellent people I work with. All these factors make the job perfect for me.

I love applying new techniques to problem solving a that mean we can better protect people from the darker side of social media – that’s what makes my job even more enjoyable and meaningful.

Of course research doesn’t always mean success. We try many ideas, some of them work, some don’t. It can be frustrating when things don't work and sometimes it's a difficult decision whether to rework and refine our method or to try something else. When research doesn’t lead to a solution for us, we can only step back, summarise, plan further and try again.

As Crisp is focused on identifying and fighting toxic content online it does mean that I see, collect and analyse difficult data content. I find it stressful at times but the team’s motivation to protect others from seeing similar content on social media makes it easier to deal with.

I like how fast things change and I enjoy being a part of that. Once humans were considered the unbeatable masters of the game ‘Go’, but then they lost for the first time to AI (which learned exactly how to win from humans). Several months later, they lost again to the second version of AI. Then the AI didn't need to learn from humans to master ‘Go’ any more - it became the undisputed master of the game instead. New ideas are coming out so frequently, that you will never feel like you knew enough about it!

If I was to give any advice to my younger self, I would say to, read more - especially about science. The world of science is endlessly interesting and I wish I had started discovering it when I was younger.

Secondly, don't get upset when things don't go as you expected, it doesn't mean it’s the end of the world - much like my subject choices in university.

Thirdly, don't stop doing things you like because you don't have encouraging teachers or parents. After I got into AI at university, I decided to pursue a Masters degree in it, but one of my tutors told me he didn't think it was a good idea because he didn't think I could do it! I did it anyway and got a distinction from Manchester University. And of course, this is now my job!

If you’re thinking about joining the industry, why not do it now and take part in building the future? Once you’ve started it, you won't regret it!

I am really glad I didn't give up. When I was at primary school in China, I was a representative in our art class until the day I drew one thing wrong and got humiliated and slapped by the art teacher in front of all the class. I stopped anything painting-related after that.

I wish I had been braver to keep going with it because I loved art and had won awards for it. I really wish that I could tell 10-year-old me that no matter what other people say, don't give up what you like. I've learned my lesson from it now and I won't let it happen again!

And to anyone wanting to join the industry, just do it as soon as possible! I truly believe AI will be the future for so many things. I think it will replace many repetitive jobs. It still has many limitations so if you’re thinking about joining the industry, why not do it now and take part in building the future? Once you’ve started it, you won't regret it!

I’d also say to keep on learning. Share learnings with colleagues and make good use of open sources - open courses like Coursera and Udacity are excellent for beginners. Even people with background knowledge should check them frequently to keep up with the latest research findings. Apart from this, reading papers from top AI general conferences such as ICML and NIPS, or application-oriented conferences like CVPR will keep you up-to-date with the latest research.

Another recommendation would be to connect with other people - AI is a big area. A quick way to exchange information is networking with other people who work in the same area, sometimes you would not believe the interesting things you've missed.

Have a story to tell that will inspire others to get into digital? Get in touch at [email protected]