As most parents know, ear infections are a common ailment among children.
“For the first three years of (my daughter’s) life, we were constantly in the pediatrician’s office,” says Kovačević. “She had ear infections all the time.”
Ear infections are treated with antibiotics, which can have negative effects on children and result in increased bacterial resistance. In 2009, there were 20 million such prescriptions in the U.S. costing $2 billion.
But ear infections are often confused with otitis media with effusion and no effusion, which presents in similar ways to an ear infection, but is not bacterial — meaning kids who do not require antibiotics are often prescribed them. So Kovačević and Dr. Alejandro Hoberman, the Chair of General Pediatrics at UPMC, teamed up to create a diagnostic aid to help doctors more accurately diagnose ear infections.
Hoberman “approached me a few years ago to see whether we could create algorithms and software that could mimic what expert otoscopists like him do when they make decisions,” Kovačević says. They developed an automated system for classifying the three diagnostic categories of otitis media that has an 80 percent accuracy rate.
Kovačević also works in “signal processing — you can think of it also as data science,” she says. She was drawn to tech as a child, when her father would present her with mathematical brain teasers (“Your mom made 10 palačinke, your brother ate three; how many are left?”)
“It has all been about patterns for me: recognizing, connecting and creating patterns,” she says. “I see structure and organization, almost like an architectural piece. I see and feel beauty in math the same way I hear and feel it in music.”
To other young women considering a career in tech, she says:
“Follow your heart and don’t be intimidated by failures; they are necessary for you to grow.”