From Observer to Leader: How Computer Vision Researcher Ilke Demir Found Her Voice

Ilke Demir ● Senior Research Scientist ● DeepScale

Ilke Demir ● Senior Research Scientist ● DeepScale

Ilke Demir is a commonly cited scholar, machine learning conference speaker, and an advocate for women in STEM. But she had no idea she would accomplish all that she has today. 

Ilke grew up in Turkey, where she liked physics and science from a young age. The first time she was introduced to computer science was when she entered into a national high school Olympiad challenge.

“That [the computer science olympiad] was the reason I was introduced to programming for the first time and I was so excited. I couldn’t believe it when I had my first ‘aha’ moment!”

During her undergraduate studies, Ilke worked in a robotics lab for an internship and was so interested in learning more about computer vision that she decided to pursue a PhD in Computer Science. Although it was an intimidating risk to take, Ilke left her friends and family in Turkey to study at Purdue University in Indiana.

During her first few years as a PhD candidate, Ilke struggled with speaking up and sharing her opinion, and usually went along with the wishes of the group. In her final year, however, Ilke became involved in a women’s network called WiSH, or Women in Shape. Ilke organized a group of WiSH women to present a Deep Learning for Geometric Shape Understanding workshop at a leading computer vision conference called CVPR. Collaborating with other women on this workshop helped Ilke grow as a researcher and as a leader.

“Before the research workshop, I was silent and in the backseat; an observer instead of a participant. But that first workshop showed me that not all collaborations should be scary. Normally those meetings rarely have people similar to me — but in that group everyone was alike and familiar, thus I felt free and confident to share my opinion. I realized I could be that version of myself in every other situation.”

Ilke originally planned to stay in academia after her PhD, but she received an exciting offer to work as a researcher at Facebook. Although it was a difficult decision, Ilke knew it was a once in a lifetime opportunity that she couldn’t turn down. Again, Ilke picked up her life and moved to California. The risk paid off, because it was at Facebook where she worked on the project that she is most proud of thus far.

“Currently about 70% of the world doesn’t have street addresses. To solve that, we applied deep learning and computer vision algorithms on satellite images … and assigned each 5x5 meter square a street address. This is applicable to developing countries with no addresses at all. According to our studies, the mailmen are now able to find the recipients and deliver the mail much faster. Seeing that real world impact was incredible.”

When it comes to advice for navigating their careers, Ilke believes that women should look back on their previous accomplishments in order to overcome current challenges. Whenever she feels like she doesn’t deserve to be where she is or starts feeling imposter syndrome, she visits her own Google scholar link as a reminder of all of her incredible achievements. 

Currently, Ilke is working at a startup called DeepScale where she develops deep learning and computer vision methods like temporal models to learn segmentation masks from videos. One of the many applications for this technology is autonomous vehicles.

“I’m excited about what deep learning and AI can do to advance computer vision.”

Ilke’s career is only just beginning, and she took risks and made bold decisions to get where she is today. Her work is at the forefront of technology and AI research, and her contributions will continue to make a difference for years to come.

This story was written by Rina Schiller, Wogrammer Journalism Fellow. Connect with her on LinkedIn, Twitter, or Instagram