Question: Tell me about something innovative you put in place, why did you do it? Why was it innovative?
Answer: I was involved in the NLP process in the company, which was a manual process. The process included the analysis of data sets for speech applications. It was too slow. I decided with a colleague to improve the process. I talked to people in the development department. I checked whether there were already tools available in the company for this task to avoid re-inventing the wheel. No tools existed. We pitched an idea, and the development of the tool started.

Q: What process steps were automated via the tool?
A: It was about getting data, analyzing language data, and annotating the data.
Q: Could you elaborate on what exactly you mean when you are saying analyzing language data and annotating data?
No data points.

A: Could you share with us what the annotation is about? In case there is an NDA, let me know.
A: Yes, there is an NDA. We use an internal algorithm that maps text directly to an intent. We do not follow the conventional approach of annotating language data first and then assign a meaning, an intent to it in a second step.

Q: Automating a manual process sounds more evolutionary than revolutionary – tell me about something you’ve done you feel was truly evolutionary?
Comment: No additional data points on innovative ideas.
A: The tool has been in production for a year. There are new features added to it, like integrating the internal algorithm for building language models.

Q: What user feedback did you get?
A: I did not get any feedback as was working in the back-end. But the tool has been in production for more than a year.

Q: What was your role in the process?
A: I provided the requirements, my proposals for improvement. I wasn’t involved in the development process of the tool.
Comment: No data points that she did any follow up during the tool development.

Analysis

Competency asserted: Invent and Simplify
Job title: Language Engineer
Interviewer role: Software Development Manager

Vote: 👎

When asked about the most innovative thing she had done, the candidate came up with an evolutionary idea, i.e., automating the NLP process steps, which had been done manually in the company.

The example was more about the Simplify in the ‘Invent and Simplify’ because of the introduction of the tool, which automated parts of the NLP process reduced the time which they needed for the manual process by 30-50%. This is her and her colleague’s guesstimate. Neither she nor anyone else collected any data to base this statement about improvement.

She did not define and use any metrics to identify the need for a change besides saying that it was too slow.

She initiated the automation together with a colleague and provided requirements. But she did not dive deep in a data-driven manner, nor did she follow up whether the change provided an improvement.

Learnings

When you prepare your stories, ask someone to challenge every fact or assumption you make. When the candidate said that the time gain was from 30% to 50%, she had no real story to back this up.

Also, notice how the interview insisted on a truly-innovative example.

Leave a Reply