Use ChatGPT for interview prep
Given a job description, use ChatGPT to anticipate potential interview questions
Job interviews are a fraught experience. One of a job applicant’s biggest fears is being asked a question for which they’re not prepared and for which they’ve no good answer. Some people are fortunate enough to be able to speak well extemporaneously. They can quickly overcome surprise at being asked an unanticipated question. But most people are not good at speaking on the spot, and so they spend inordinate amounts of time preparing for each interview, trying to divine questions that they may be asked, and having good stories to relate career accomplishments to the job for which they’re interviewing.
And, frankly, a lot of this time is probably poorly spent, because the interviewee has no way of knowing ahead of time whether the questions she practice on will actually be asked in the interview. What if there were a different way? What if you could just put the job description into ChatGPT and ask it to generate some likely questions?
Then you can spend your time coming up with plausible answers for those questions. And since you can repeatedly ask ChatGPT to come up with possible questions, you can very quickly come up with dozens. Figuring out how to relate all of these questions to your career accomplishments, and how those accomplishments inform your ability to do the job for which you’re being interviewed, sounds like a better use of your time.
Thanks for reading Buy the Rumor; Sell the News! Subscribe for free to receive new posts and support my work.
So, in a bit of meta commentary on ChatGPT, jobs, and prompt engineering, there’s this job post, for a Prompt Engineer and Librarian at Anthropic AI. So I created a prompt, asking ChatGPT to come up with some relevant questions: Consider this job description, and come up with 10 questions likely to be asked in the interview. I pasted the text of the job post after the prompt, and hit enter.
Here’s what it returned:
Here are 10 questions likely to be asked in the interview for the Prompt Engineer and Librarian position:
Can you tell us about your experience with prompt engineering for large language models or image generation models?
How do you stay up-to-date with emerging research and industry trends in the field of AI?
Can you walk us through your process for discovering, testing, and documenting best practices for a new task?
How would you prioritize tasks and determine which ones are most relevant to our customers?
Can you give an example of a complex behavior you were able to achieve using a series of well-crafted prompts?
How do you approach creating documentation and tutorials for technical concepts?
Have you built up a library of high-quality prompts or prompt chains before? If so, can you tell us about that experience?
Can you explain how you would go about teaching the art of prompt engineering to our customers?
How do you anticipate and mitigate unforeseen risks associated with new technologies?
How do you collaborate with others, particularly those with different areas of expertise, to build effective teams and solve complex problems?
Then I asked it to answer some of its questions. Here’s its answer for Question 1:
Certainly, I have experience with prompt engineering for large language models, and I'm excited to apply my skills to this role. In my previous job, I worked on developing a chatbot for a customer service company that used GPT-3 as the underlying language model. To make the chatbot more effective, I experimented with different types of prompts and prompt chains to elicit the desired responses from the model. I found that by using carefully crafted prompts, I could guide the model to provide more accurate and helpful responses to customers. Additionally, I have worked on image generation models and have used prompt engineering techniques to generate specific types of images. I have also kept up to date with recent research in this area and am excited to apply these learnings to new challenges in this position.
Now, this answer isn’t great, in that it doesn’t tie back to the interviewee’s career accomplishments, and how those demonstrate her ability to do this job. However, it is sufficient to get the interviewee thinking about her own answer to this question.
The answer for Question 2 is equally problematic, in that it doesn’t tie to the interviewee’s career accomplishments, and how those are relevant to the job for which she is interviewing, but again, it will help her start to think about how to craft a better answer:
Staying up-to-date with emerging research and industry trends is an essential aspect of being an effective prompt engineer and librarian. To stay informed, I make it a point to read academic papers, attend conferences, and follow industry thought leaders on social media. I also stay engaged with online communities focused on prompt engineering, where experts share the latest insights and best practices. I believe that being involved in these communities helps me stay current on new developments and technologies, and it also provides a great opportunity to exchange ideas with other professionals in the field. In addition, I am always seeking out opportunities to expand my knowledge by taking online courses, participating in workshops, and engaging in other forms of professional development. Overall, I believe that staying informed and up-to-date is critical to being an effective prompt engineer and librarian.
This is a great example of how ChatGPT can be used for divergent thinking. And since the nature of prompts means that you can tweak them, and re-run them, you can quickly generate dozens of ideas. And each idea can be explored in further detail. And prompts allow this to be all done very quickly, repeatably and scalably. That’s the beauty of the ChatGPT interface: you can quickly generate dozens or hundreds of ideas from simple prompts. It is a highly scalable adjunct to your brain.