LEL alumni: Shilin Gao

Posted on February 3, 2021 by

Shilin Gao graduated from LEL in 2018 with a degree in Linguistics and Social Anthropology. She then completed an MSc in Speech and Language Processing at the University of Edinburgh, and she now works as a speech recognition engineer at Speechmatics in Cambridge. 

Can you tell us more about your job? 

The work is mostly about maintaining and developing speech recognition systems. Our company provides speech recognition technology to a wide range of customers like media companies that would use our technology to generate transcriptions for the videos, or call centres or other institutions that want to monitor the customer calls for regulation reasons.  

Initially, my job was to build language packs for new languages including Turkish and Malay. More recently, I’ve been working on some product features that improves the recognition accuracy for customers’ specific use case.

Do you have any advice for Linguistics students who’d like to follow in your footsteps and work in speech tech? 

Assuming that they have already developed interest in speech technology there are two things they can do to prepare themselves. The first thing is to take some courses in mathematics and statistics. At Manchester, students have 20 credits a year for optional courses. I spent those credits studying Further Mathematics and Further Statistics with the School of Social Sciences. The other thing is programming. I joined the Code Club at the University of Manchester which gave me a chance to learn Python. Then there were also opportunities in linguistic courses, for instance in the Quantitative Methods in Language Sciences we used the programming language R, which gave me experience in combining statistical problems programming exercises combined with language data; Forensic Linguistics was another course in which we used programming methods to analyse textual data: both are very useful in preparing for work in speech and language technology.

Dissertation is a great opportunity to focus on a more tech-y topic in a relevant field, for example, natural language processing (NLP), speech recognition and speech synthesis. My dissertation project was “An Evaluation of Parts-of-speech Features in Native Language Identification (NLI) Tasks”. NLI is a subfield of NLP [Natural Language Processing] that uses computational techniques to predict the author’s native language based on an anonymous piece of writing. I learned so much from this experience and it also helped with my application for a relevant Master degree.

In summary, the opportunities are there in the curriculum but it’s up to the students to find them and use them. For an undergraduate student in linguistics or social science, it’s very hard to immediately get hands-on experience of real speech technology, but it’s good to practice solving problems which involve programming and make use of speech or text data.

Do you have any other more general advice that students could use in preparing for their career post-graduation? 

Looking back at my own student experience, one thing that was really helpful – and I never regret doing that – is that I tried a lot of options career options during my time at university, which helped me find out what I really wanted to do. As soon as I arrived in Manchester, I started looking for part-time jobs and internships using the excellent Careers Service on campus. It took me six months to get my first offer, where I worked as a customer service assistant in Lloyds Bank in the Manchester city centre for more than a year. It’s good to start early because it can take a long time and lots of trial and error to find your first job.

Starting early also gives you more time to try out different things. When you apply for a degree like Linguistics and Social Anthropology, you don’t necessarily think that you’ll end up working in speech technology. I considered many different options, and by the time I applied for postgraduate degrees I still couldn’t make a choice between the tech industry/programming and the finance and business path. So I applied for programmes in both areas and kept exploring relevant work experience. For the finance/business area, apart from my job for the Lloyds Bank, I also did an internship in the investment banking equity research sector in China. On the tech side, I did a Q-Step internship with a Manchester-based tech company, Reason Digital. The exposure to different work content and workplace atmosphere made me feel that programming work was a better path for me. Hands-on work experience really helped me to make an informed decision about my career, because having explored and compared the options, I knew I made the right choice and there is nothing to regret.

The featured image is of Shilin presenting a poster on her undergraduate dissertation.

Manchester Q-Step offers internship opportunities in data-driven roles to undergraduate students, including students in Linguistics and English Language. Internships for 2021 will be announced imminently. Eligible students will receive an email once new opportunities have become available.

The Code Club is now online – you can find out more about it here.

Posted in: alumni, jobs