These two job seekers built an AI chatbot to talk to recruiters

Joshua Curry and Vishal Patil have seen a lot of customer service chatbots. Chat windows that pop up on your screen while visiting a variety of sites, from online retailers to mobile phone carriers, asking for help have proliferated in recent years.

Curry and Patil built a similar chatbot. However, their services do not exist on business websites or solve customer service issues. Their work is featured on a portfolio website and is intended to help them find work.

Curry and Patil’s personalized AI chatbot uses your application and professional experience to interact with recruiters visiting your website. They hope chatbots will help them land jobs in today’s tough “low-employment” job market. U.S. employers added just 116,000 jobs last year, with 1.46 million jobs in 2024. Their approach is just one of many creative solutions applicants are experimenting with in hopes of getting noticed, including sending snail mail and relying on reverse recruitment agencies.

Curry, a web developer in San Francisco, has interviewed for two of the approximately 140 positions he has applied for since November. “That’s insane,” he says. “You need to stand out.”

Patil, an F-1 student visa holder, has found the job market particularly difficult, he said. He estimates he received 30 selection calls or interviews out of more than 700 applications he submitted from mid-August to March 2025.

Job seekers today face stiff competition. Approximately 8,200 applications are submitted every minute to LinkedIn, and 38% of U.S. job seekers are applying to more jobs than ever before, but receiving fewer responses, the company told CNBC Make It in April.

Patil originally built the chatbot VAi to solve a persistent problem he faced when searching for summer internships. The 26-year-old was juggling classes for a master’s degree in software engineering at the University of Maryland, College Park with an on-campus job in IT, which meant he wasn’t able to respond to recruiters as quickly as he would have liked. He felt that even the slightest delay would “impede” the search.

“They have thousands of applications,” Patil said, adding that he felt “if we don’t get the information soon, they’ll move on to the next candidate.” He thought it needed to be faster, allowing chatbots to respond instantly.

Curry created the chatbot ChatJC after seeing several job postings looking for people to build something similar. He wanted to show that he could be that kind of person.

Hiring managers in his field of work “want to read the code you write.” Releasing ChatJC as an open source project “allows people to see my work,” he says, and allows others to create their own applicant chatbots.

Plus, “I like the idea of ​​having something do the work for you while you sleep,” he says.

Both Patil and Currie say it took about two weeks to start designing the chatbot and deploying it on the website.

VAi uses information from Putil’s LinkedIn, resume, and portfolio website. You can start conversations with suggested prompts such as “Give me a brief overview of Vishal” or “What is Vishal’s strongest skill?” ChatJC similarly utilizes materials such as Curry’s resume, cover letters, testimonials, and volunteer work.

There were some problems at first. For example, on ChatJC, some visitors typed profanity into the question field. Patil and Curry created guardrails to prevent chatbots from providing sensitive personal information, such as home addresses, or answering irrelevant questions.

There is also the issue of hallucinations. Mr. Currie instructed ChatJC to only state what it found from the materials provided. But as he points out, and as LLM shows, “chatbots don’t always follow that direction.” Patil similarly tells the chatbot to “stay on track” and “don’t lie” about its experience, such as by claiming it has skills it doesn’t have.

Callie can see the questions people ask, but she doesn’t know who will ask them. So far, questions include “What is his tech stack?” (ChatJC answered well, Curry said.) “What is Joshua’s main weakness?” (ChatJC did not answer.) “Can he code like crazy?” According to his analysis, the number of unique visitors since launching in March was 80.

According to Patil’s analysis, VAi has received more than 3,300 views since its January launch. He says he had 328 unique visits and 492 questions in the last 30 days, but he doesn’t know what those questions were.

Patil says writing and posting about VAi has increased his visibility and reach on LinkedIn, and brought attention to his profile during his job search. “No it [the chatbot]“I don’t think that happened, and it opened up a lot of opportunities for me,” he says.

Patil added that a small number of recruiters, such as those for which he applied, gave him praise and feedback on the VAi, and each offered him either an interview or a technical evaluation. Several colleagues asked how he built VAi, and one offered to introduce Patil to the job.

Before joining VAi, he says, “I had never had this discussion with a colleague or a recruiter.” “I was struggling all by myself.”

“People who are actually applying for jobs get lost in the shuffle,” Curry said. As a candidate, he added, “you just send this stuff out into the void and nothing comes back shiny.”

ChatJC is his answer of sorts to the use of AI in the hiring process by employers. “I don’t think it’s a game.” [the system]. “I see it as consistent,” Curry said. By “utilizing the same tools” on the applicant side, “the goal is to be on equal footing,” he said, adding, “It’s really just elevating ourselves and saying, ‘Look, we’re not just rats in a cage running through a maze.'”

Patil secured a summer internship. He says he was offered his current job in the university’s IT department in part because his colleagues were impressed with VAi.

For Curry, so far ChatJC has been a great conversation starter in the tech world and provided useful practice for building chatbots, he says. But as with his other search efforts, his larger purpose is clear. “I hope this damn thing gets me a job.”

Do you want to get ahead at work? Next, you need to learn how to make effective small talk. In CNBC’s new online course, how to talk to people at workour expert instructors share practical strategies to help you use everyday conversations to increase visibility, build meaningful relationships, and accelerate career growth. Sign up now! Use coupon code EARLYBIRD for 20% off. Offer valid from April 20, 2026 to May 4, 2026. Terms and conditions apply.