ZipRecruiter: Vision Sprint

I had the privilege of leading ZipRecruiter's product design team in an ambitious design sprint focused on reimagining the future of finding work. As the sprint leader, I set the strategic direction, worked closely with each designer, and orchestrated the final presentation delivered to our executive leadership, including the SVPs of Product, COO, and CEO.

Our Vision Sprint challenged the status quo, proposing a shift from our traditional ad-based model to a state-of-the-art matchmaking experience. Our new approach addressed critical pain points for employers and job seekers in our marketplace.

What follows is a comprehensive case study showcasing the innovative work of our eight-person design team. It represents the culmination of this intensive sprint and encapsulates insights from my twelve-year tenure at ZipRecruiter.

The process looks something like this: It's a manual process with too many steps and much back-and-forth, often leading to frustration for both participating sides.

With natural language processing we can dynamically understand exactly what job seekers care about and use those vectors in our matching.

Using LLM embeddings, we can understand you in a holistic and multi-dimensional way that today's structured data can't attain.

Simply answer some questions about your qualifications, our voice AI will turn this conversation into a robust profile that showcases your unique experience.

You don't have to be good at this. We'll assist you every step with every question. After the system understands you, it will represent you perfectly.

Jobs aren't this boring. The company, the role, the team, etc. are more interesting than a wall of text. Job seekers want to know what it's really like to work there.

Wouldn't it be great to showcase the role in a way that gets job seekers excited about the opportunity?

Let's face it, this absolutely sucks! Who wants to do this over and over again just to not hear back from 100 different employers? Pain.

And dude... trust me, I've reviewed 10,000 applications in my time. As a hiring manager, there are countless other things I could prioritize.

Getting ready to talk to an employer, if you're preparing which you should, takes time.

Our fine-tuned LLM can help you prepare for interviews with specific companies. You can even switch to voice mode to practice in a real-life situation.

From our user research, job seekers often have a difficult time expressing who they are in an interview situation.

No worries, we'll assist during the call to remind you to mention key aspects of your qualifications that are specific to the interviewer's question.

The Vision Sprint was well-received by leadership. We implemented cutting-edge features like LLM-based preference collection in our eLLMo project. However, the high-level experience of no search and no applications was rejected due to our reliance on ad-based revenue.

Before leaving Zip, I warned that failing to disrupt our own business model could lead to falling behind. I drew a parallel to Kodak, who failed to embrace digital cameras to protect their film sales, ultimately leading to their downfall. This cautionary tale emphasizes the critical need to innovate, even at the risk of disrupting our current revenue streams.