For seven weeks this summer, we, Luke Zermani and Aidan Zheng, were given the chance to work for Nuance Communications. Nuance uses artificial intelligence to automate human speech recognition and improve customer experiences across various industries. Nuance’s software is applied in industries like finance, telecommunication, retail, healthcare, and government. We worked with part of Nuance’s medical team whose primary goal was to develop the software for doctors in hospitals. Through our time working at Nuance, some features of the healthcare model that we observed were ordering medicine, keeping and calling records, and communicating with staff and patients.
During our seven weeks, we were assigned a large variety of tasks, most of which were centered around testing the voice recognition software and trying to find errors. We would be given a specific list of phrases often used in the medical field, and we would test them by saying these phrases into our software. When the software did not return the output that we expected, we often analyzed the WAV files to try to determine the cause of the error. Oftentimes this error would be caused by a minor hiccup in our pronunciation, so in that case, we would simply retest the phrase until we would regularly receive the correct response. However, sometimes the error would be caused by a problem with the software, and in that case, we would report this to a more experienced software engineer to figure out the error.
The team operated in a workflow called a “scrum”, which is part of a type of larger framework called “agile” that is known for incremental development. Scrum, more specifically, breaks down development into chunks called “sprints''. At Nuance, it was common for a sprint to last two weeks. Before the start of a sprint, the team had a meeting called “sprint planning” where a list of tasks is created. During the sprint, we had daily meetings, called “daily scrum” at Nuance, in which each team member shared their plans and progress through the day. These meetings were mediated by the scrum master, Yolande, who helped keep each meeting focused and organized. A “sprint retrospective” meeting was held at the end of sprints as a time for the team to reflect on the sprint and identify what went well and what could be improved. It was very interesting to learn about this type of workflow that large companies use and strategies to stay productive and efficient when working on a project as large as Nuance does.
Testing medical phrases using the software Writing code to help organize audio files
Our workflow revolved around Microsoft Teams, an application that included communication in the form of messaging and virtual meetings. It also had a calendar to keep track of and create meetings and areas for projects and questions that we had. Teams made the workflow during the day very efficient. It made it easy to ask a specific team member a question or set up impromptu meetings. Another application we used was an application called “Jira” to keep track of the progress that we make on any given task. Jira is used by all of the full-time employees as well, so they can see where we stand on a certain assignment we are given. Because Nuance is mostly a virtual company, using Teams was necessary to help everyone communicate and work from home. While in-person meetings were rare, they were often reserved for team-building events like getting ice cream together.
We would also participate in Nuance’s intern career development series meetings, where a full-time Nuance employee would give a presentation to the interns in the US, Canada, and India. There were also many in-person events in the Burlington area. There was a trash clean-up day, and there were multiple days when the interns and their teams would go out for ice cream. We had a great time this summer, and thanks to our manager Matt Harrison for helping guide us through such an amazing experience this summer.
Comments