Discover insights on OpenAI GPT-3 from Oras Al-Kubaisi, founder of 'Definite Guide to Building Products with OpenAI' course. Expert tips and strategies revealed.
This lesson is from Definite Guide to building products with OpenAI.
Job Description ai is a job description generator software that uses OpenAI GPT-3 API to generate high-quality professional job descriptions for businesses attracting the best talent. Screenshot 2021-02-22 at 10.54.37.png 222.53 KB Who are you and what is your backstory?
My name is Oras Al-Kubaisi and I’ve been a software engineer for more than 15 years.
What have you built with OpenAI and what does it do?
I built Job Description AI, a service to generate professional job descriptions in a few clicks.
How did you come up with the idea for your product?
I have started a few months ago working on an ATS (Applicant Tracking System) to help startups avoid hiring mistakes by introducing a process that will create the job description, create a hiring pipeline, parse CVs and even show some examples of what to ask and how to prepare for the interview as an interviewer.
My original idea for generating job descriptions was using templates based on the job titles.
While brainstorming, I thought what if I try generating job descriptions using GPT-3? Long story short, I received my access around mid-December and started playing around with GPT-3 and was really impressed by it.
At this point, I thought, ok instead of waiting for months until I finish the ATS; how about creating this tiny service just to generate job descriptions?
I was lucky to find jobdescription.ai is available! So I registered the domain and started building. Three weeks later I finished the app and launched it on Product Hunt.
Why did you decide to build with OpenAI API? Have you had other exposure to AI tech in the past?
Back in 2018, I started working with some machine learning projects inspired by fast.ai videos. I joined Kaggle and started learning about binary classifiers for images. Later on, I started reading about NLP and all libraries related to them.
How is your product doing today, financially and what does the future look like to you?
I only started working on it full time on the 1st of February. It’s been 3 weeks now and there is a good attraction. I have joined the YC Startup School sprint with a goal to secure a pilot for the first client. The coming 3 weeks will determine how I performed to achieve this goal.
Are you worried about Platform risk? What are you doing to mitigate that?
Not much, I have followed the safety best practices religiously and I even added another layer of check to ensure the data prompt itself is safe.
What are your long-term plans for the product? Is your product self-funded or have you received Angel/VC funding? Are you planning to raise VC funding?
The product is self-funded for now and based on the growth I might raise VC funding but before committing to that I would like to prove there is value in using the service. I am discovering the best way to add value to customers.
What is your advice for founders looking to build products with OpenAI API? Should first-time founders be building their first product build with OpenAI?
They should, and my advice is: Just start! The playground is easy to work with and there are lots of resources and tutorials inside the platform to help you start building. Being the first product does not matter here, but I would say expect longer delays. There are lots of things to consider:
Good idea.
Safety measures (it could take up to 3 weeks to implement safety measures).
Pre-Launch review. For any new feature you will need to submit a long-form explaining the feature, safety measures added, mitigation process … etc.
Is building an AI-enabled product any different than building without AI? What are the pitfalls to avoid while building with OpenAI?
Few things:
The latency! We are used to faster products by now and having to wait several minutes to get a response is not something the user is comfortable with. Add to that handling errors from the OpenAI side and monitoring the service is responding.
The cost. You are going to pay per request and depending on which engine you go for and the prompt length, this could become costly in no time.
What kind of learning resources(books, podcasts, etc) you would recommend to the founders looking to build AI-enabled products?
I don’t know if there is a book at all but I would recommend the OpenAI slack channel. This could be a good opportunity to write a book (build 10 real products using OpenAI :) ).
Where can we go to learn more about you and your product?
I am writing this guide in public and making it available for reading online.
This guide is also available in formats suitable for Book Reading Apps and Devices(PDF, ePub, Kindle). If you like and want to support this work, you can Pre-Order it here.
