The Future of Hiring: With Nir Dovrat

Insights from Nir Dovrat , Co-Founder and CTO of Canditech

This week in our spotlight on Israeli innovation, I had the pleasure of speaking with Nir Dovrat , Co-Founder and CTO of Canditech . Nir and his team are doing fascinating work with their AI-powered candidate evaluation platform, which combines tools like job simulations, cognitive and personality assessments, video interviews, and pre-screening chatbots.

Your help is needed: Take a look at this new format featuring an in-depth interview, aiming to provide deep insights into the minds of innovators. Do you enjoy diving deep? Let me know in the comments!

Q: Nir, what inspired you to create Canditech?

The inspiration came from my own challenges as a CEO in my previous company, which grew from 4 to over 40 employees in just a year. I had to recruit for various roles from tech to non-tech, both at junior and senior levels. For some positions, I received 50 CVs, while for others, I got 400. This process highlighted how inefficient traditional hiring practices can be, largely because they rely so heavily on CVs and job descriptions, which led my managers and myself to spend so much time reviewing resumes, while overlooking great candidates

It became clear to me how arbitrary and restrictive the process often is. For example, requiring "4 years of experience" can exclude outstanding candidates with just 2-3 years who could excel in the role. And does attending a prestigious university automatically make someone a better employee? Of course not. I found myself searching for tools to evaluate candidates beyond resumes but discovered that the available solutions were lacking. Platforms designed to test specific skills like coding didn’t fully assess a developer’s abilities, such as communication, problem-solving, or design thinking. And for non-technical roles, I couldn't find tools that I believe are unbiased, valid and will provide the candidate experience I was aiming for.

After two years of research and speaking with HR leaders, hiring managers and psychometric experts, we learned that the most effective predictor of job success is realistic job simulations: tests that mimic the day-to-day challenges of a given role. This approach is a win-win for both hiring teams and candidates. Work sample tests are not only considered the most reliable tool for evaluating candidates, but they also give candidates an early glimpse of their potential future job, helping them determine their interest in the role from the outset. This concept forms the foundation of Canditech: a platform that enables hiring teams to create customized, realistic job simulations for any role, delivering a more accurate, fair, and engaging way to evaluate talent.

Q: With AI playing such a big role in recruitment, there’s a lot of talk about bias in algorithms. How do you make sure Canditech stays fair?

That’s indeed one of the greatest challenges in AI-driven recruitment. At Canditech, we’ve found that combining deterministic scoring algorithms with AI strikes the best balance to keep bias in check. This hybrid approach ensures fairness by grounding the AI's recommendations in clear, measurable criteria while leveraging AI's strengths to evaluate complex patterns.

Developing this system took over a year of research and experimentation to find the ideal implementation. We discovered a unique method using a robust multi-agent model, where each agent focuses on a specific segment of the assessment. This modular approach allows us to minimize bias by breaking down the evaluation process into smaller, objective components. We’ve tested this extensively on thousands of answers and conduct regular audits to ensure its integrity and fairness.

Another critical aspect of ethical AI in hiring is avoiding the “black box” problem. Transparency is key: not just for meeting regulations but also for building trust with HR teams and hiring managers. Our system provides detailed insights into why a candidate received a specific score, enabling hiring teams to understand and even refine the criteria to better align with their unique needs. This transparency ensures our platform remains fair, adaptable, and compliant while fostering confidence among users.

Q: You provide a lot of metrics and data, but how do you make sure the human element isn’t lost?

The beauty of our platform lies in how it encourages managers to engage directly with the evaluation process, ensuring that the human element remains central to hiring decisions. Canditech recommends which candidates to focus on based on their performance, but before proceeding with a 1-2 hour interview process, hiring managers can spend just 1-2 minutes reviewing a candidate’s results. This allows them to see firsthand how candidates approached challenges, explained their thought processes, and demonstrated their potential.

Our goal isn’t to replace the human touch but to enhance it by giving more candidates a fair opportunity to showcase their skills. This approach has been transformative. Hiring managers frequently share how they’ve discovered incredible employees they would overlook in a traditional process.

The results speak for themselves. Many of our clients report “positively bizarre” scenarios where their hiring timelines for developers, salespeople, or marketing professionals have dropped from months to just 1-2 weeks. This balance of data-driven insights with human judgment has proven to be a game-changer, leading to faster, more effective, and inclusive hiring outcomes.

Q: HR professionals are quite cautious about adopting AI-driven solutions, especially when it comes to something as critical as hiring. How have you approached showing recruiters and candidates the value of Canditech?

A:That’s a great question and one we’ve navigated throughout our journey. The HR landscape is undergoing a significant transformation, driven in part by the rise of AI. HR professionals are increasingly recognizing that to meet the growing and dynamic needs of their organizations, they must embrace technology, particularly AI-driven solutions. Over the past year, we’ve seen a marked shift, with more and more HR professionals proactively reaching out to Canditech to explore how our platform can streamline their hiring processes.

What also sets us apart is that while Canditech is managed by HR, it’s ultimately a platform designed for hiring teams. Take the sales team, for example - if they’re used to hiring salespeople through a process that drags on for 2-3 months, and they discover that with Canditech they can achieve the same results in just 2-4 weeks, it’s often the sales team themselves that become our biggest advocates. This alignment of efficiency with tangible results resonates deeply across all levels of an organization, demonstrating our value not just to HR but to everyone involved in the hiring process.

Q: As AI becomes more advanced, the potential for misuse grows. For instance, candidates could exploit vulnerabilities like prompt injection or other adversarial techniques to manipulate their scores. How can innovators build systems that effectively detect and defend against such attacks?

A:You’ve touched on a fascinating and ongoing challenge at Canditech - how to stay ahead of candidates who might try to game the system. Interestingly, cheating is much less common than many expect, but it’s still something we proactively address. It’s similar to the virus-versus-antivirus dynamic: every potential exploit requires a smart and robust countermeasure.

We’ve implemented a wide range of tools to detect anomalies during assessments. These include monitoring window activity, detecting copy-paste attempts, analyzing typing speed, tracking cursor movement, and much more. For more advanced tactics like prompt injections, we’ve developed unique solutions using some of our "agent" models to specifically identify and neutralize these types of exploits. While it’s not the primary challenge we face, it’s certainly one of the first things we considered when incorporating AI into interviews.

For us, this is one of the fun aspects of the job: getting creative and thinking like a potential cheater so we can build intelligent ways to prevent and detect it. The result is that cheating on Canditech is incredibly difficult without being caught. And even if someone tries, the risk-reward simply isn’t worth it. It’s akin to lying on your resume: you might get an interview, but you’ll likely get rejected immediately once the truth comes out. Our systems ensure that integrity remains a cornerstone of the hiring process.

Q: Where do you see hiring going in the next 5–10 years?

A:I believe that in the near future, AI will become an even more interactive part of our lives, touching nearly every aspect of how we work, including HR. The future of candidate evaluations and interviews lies in leveraging AI to create personalized, interactive assessments. These will not only evaluate candidates for a role but also help uncover their true potential in a way that static, traditional methods simply can’t.

At Canditech, we’re already working on groundbreaking innovations in this space, including an AI interviewer powered by our multi-agent technique. This system will truly understand candidates’ answers, ask insightful follow-up questions, and provide a seamless, engaging experience. Imagine an AI capable of conducting thousands of interviews per day and learning from every single one. How long would it take for such a system to become one of the best interviewers in your company? The potential here is enormous—not just for efficiency but also for making smarter, more informed hiring decisions.

Our long-term vision is to redefine how organizations and candidates connect, ensuring the process is not only efficient and fair but also deeply insightful and tailored. We see a future where hiring isn’t just about filling roles but about unlocking human potential at scale.

Q: What’s something you’ve learned along the way that others should know?

A: My advice to entrepreneurs tackling sensitive areas like hiring is to always start with the user and their challenges, not the technology. It’s easy to get caught up in the excitement of AI, but the real value lies in understanding the problem deeply and designing a solution that fits seamlessly into people’s workflows. At Canditech, we learned that building trust - through transparency, fairness, and usability - is just as important as the tech itself.

One of the key lessons I’ve learned is that simplicity often beats complexity. AI can be powerful, but it’s most effective when it complements rather than overwhelms. Instead of trying to automate everything, focus on enhancing human judgment and creating tools that feel intuitive and actionable. After all, technology should serve the people using it, not the other way around.