Bardia Azizian

MASc '23

What was a highlight or favourite moment from your time at SFU?
One of the highlights of my time at SFU was the exhilarating experience of finally meeting my supervisor and research group in person. Due to the COVID situation, our initial meetings were held online, and I never had the opportunity to connect with my collaborators face-to-face until the last semesters of my program. This experience was my favorite because it brought a sense of belonging for the first time since I moved to Vancouver.

How has your experience at SFU prepared you for your career path?
Before SFU, I hadn’t delved deep into machine learning and its real-world applications. However, working alongside an incredible research team at the Multimedia Lab and taking relevant courses allowed me to explore practical technologies within my research. This hands-on experience complemented my theoretical knowledge and equipped me with skills that will undoubtedly contribute to my future career success.

What advice would you like to share to students in their first year?
I would like to encourage new students to actively engage in social activities and networking. Building connections and forming friendships can greatly shape your personality and open doors to opportunities and fresh ideas for your future career. Remember, it's not just about your studies; make sure to have fun and enjoy every aspect of your life as well. 

What are your current plans?
I have recently begun my Ph.D. studies at SFU and plan to continue my research with the same supervisor. I aim to enhance and expand upon the work I accomplished during my master's program by incorporating new ideas.

What was your research focus?
My research primarily focused on visual coding for machines, addressing important challenges such as privacy concerns. Within this field, my objective was to develop efficient and privacy-preserving methods for compressing visual data for machine vision models. Tasks such as image classification, object detection, etc., require specialized codecs designed specifically for these tasks, as relying solely on human-targeted codecs may not yield optimal outcomes.