Interview with Soumil – Founder of SkillPool

SkillPool is a hiring solution that simplifies the recruitment process. If you’ve ever felt overwhelmed by a large number of job applicants, SkillPool uses advanced AI resume screening technology to assist you in identifying the ideal candidates efficiently. The platform offers a straightforward 3-step process that streamlines the screening of your applicant pool. By leveraging state-of-the-art technology, SkillPool aims to minimize the time and effort you spend on finding qualified candidates, ensuring a more efficient and effective hiring process.

Chris: Hi there, please introduce yourself.

Soumil: Hello! My name is Soumil Rathi, and I am the founder of SkillPool. Currently, I study Computer Science at the University of Illinois Urbana-Champaign.

I’ve been a fan of Artificial Intelligence since I was young, and can remember always wanting to learn more about the technology. This has led me to finding a variety of ways to experiment and create with this technology, eventually leading me to establish SkillPool.

Before I worked on SkillPool, I used to offer freelancing services in creating Artificial Intelligence models, which is also how I funded the early capital requirements for this startup. However, my passion for building with AI started long before that, all the way from when I built various AI projects for accessibility.

For the past 5 years, I have been working on improving my AI skills, as well as focusing on practically implementing those skills. This passion, combined with my desire for entrepreneurship and building is what led to me founding SkillPool.

Chris: SkillPool is designed to automate the resume screening process. Could you explain how the AI engine works to scan resumes for skills and keywords, and what criteria are used to determine a candidate’s match to a job opening?

Soumil: SkillPool is an AI-based resume screening software, and so, our main focus is on the AI we use to scan the resumes as well as the job descriptions. One of the important things about our AI is that it doesn’t actually scan the resumes or the job descriptions for skills and keywords. Instead, we use AI to scan the entire document and contextualize it within a numerical format.

The way we do this is through a type of AI model called the transformer model. Such models are extremely common, and are currently used to power Large Language Models (LLMs) such as ChatGPT. Now, while I don’t want to go into too much detail regarding the specific functioning and calculations within a transformer model, there are many benefits to scanning documents this way, as opposed to simply keyword searching.

The transformer model manages to note the complete context of the documents it scans, and ensures that the intention and meaning of the document is conveyed in the numerical and vectorised implementation.

We use a transformer model to scan the job description as well as each resume. Then, we directly compare the numerical versions of both documents, giving us a similarity score, which we then use to rank the resumes for the job opening.

Chris: The ability to rank candidates based on their match to a job opening is a valuable feature. How does SkillPool calculate this ranking, and how can recruiters best use these rankings for the specific needs of their organization?

Soumil: The ranking of the candidates is done majorly using our AI model. Once the model creates the numerical representations of both the resumes and the job descriptions, the software then compares those representations. This comparison is done particularly using cosine similarity, which is a method used to compare two vectors (the numerical format of the documents).

Intuitively, the cosine similarity tells you how ‘closely related’ two vectors are. In this case, the cosine similarity indicates how suited the resumes are to the job description. Thus, this value is then used as a ranking for the resumes. We have found that using this the cosine similarity as a ranking factor has been effective in identifying the resumes best suited for each job.

We see this ranking being used as a tool to screen out the resumes which are unsuited for the job, instead of automatically selecting the resume with the highest score. SkillPool helps make the initial job screening quick and automatic, ensuring that the applicants who are suited for the job can proceed with the application process in a far more efficient manner, allowing the recruitment to be streamlined.

Chris: Many recruiters may have concerns that such tools can harm their job security. Considering your claim that SkillPool will make the recruiting process more efficient than before, what are your thoughts on that matter?

Soumil: I understand such concerns, of course, especially considering that many people are already in the process of losing their jobs because of AI solutions. However, I would like to assure recruiters that at SkillPool, we don’t intend to be an AI software that displaces the jobs of humans. Instead, we have specifically designed our software to perform best in hybrid with recruiters, improving the productivity and efficiency of the recruitment process.

At SkillPool, we believe that the AI models aren’t capable of making decisions based on behavioral and moral traits, thus requiring recruiters to move forward and decide between the most promising candidates. Thus, we don’t pretend to find recruiters the best candidates, we simply screen those unsuited for the particular job.

As such, we make it so that, instead of being hesitant about SkillPool, recruiters can use the platform to streamline their process and greatly increase their efficiency.

Chris: Recruiting agencies, in-house teams, and small businesses can all benefit from SkillPool. Can you share examples of how each of these user groups can use the platform to improve their hiring processes?

Soumil: At SkillPool, we’ve seen users come from all types of organizations, including recruiting teams, small businesses, as well as individual recruiters. Of course, each of these users have different needs. Some need to only hire a few applicants, while others have to hire multiple applicants across a number of jobs.

SkillPool has different options and plans for each of these types of users. Whether you only want to screen through ten resumes, or whether you wish to screen through hundreds, SkillPool is the solution you need.

Chris: What sets SkillPool apart from other resume screening solutions, and how do its features and capabilities compare to alternative platforms in the market?

Soumil: Of course, there are a great number of competitors in this sector. The majority of our competitors, however, focus primarily on solutions implementing keyword matching to determine the similarity between the job description and resumes. As keyword matching fails to capture the semantic context of each resume, it is likely that such solutions can be inaccurate, leading to unsuited resumes being ranked high, or worse, suited applicants failing to pass through the screening process.

SkillPool, on the other hand, offers an improved solution by implementing advanced AI technology to get a complete and comprehensive overview of the job as well as of each candidate. Thus,we provide a more accurate resume screening solution as compared to our competitors, while also ensuring that all users of SkillPool get a pleasurable user experience, setting us apart from other resume screening solutions.

Chris: As SkillPool continues to evolve, are there any upcoming features or developments that users can anticipate, and how will these enhance the user experience and effectiveness of the platform?

Soumil: At SkillPool, we constantly seek to improve our software and offer the best possible experience to all our users. As a result, we always seek out user feedback to identify upcoming features and developments. One of the features we intend to add in the near future is the option for recruiters of the same organization to work in teams, enabling lots of collaborative features during the recruitment process. We believe adding the option of working collaboratively on SkillPool will be highly beneficial for recruitment teams.

In addition, we also want to add the option for users to directly import job postings and applicants from popular job boards such as Indeed and LinkedIn. We want to make the recruitment process extremely streamlined and convenient for all recruiters, and directly importing data from job boards will go a long way in achieving this goal.

Overall, we are devoted to constantly improving our product and experience, and will be releasing regular updates and new features to the platform.

Chris: Can you share some examples of industries or job roles where SkillPool has proven particularly effective in identifying the best candidates, and how it adapts to different sectors and positions?

Soumil: Throughout the development of SkillPool, we wanted to ensure that there are no restrictions on SkillPool’s accuracy and performance across all industries. In fact, this is another one of the advantages of using an AI model like ours as compared to a keyword matching model. Since our model is based on deriving the context of the job description and resumes, it remains unbiased across all industries and job roles, and performs equally as good in all of them.

Since the product was released, we have observed organizations across a large variety of industries using SkillPool to accurately and efficiently screen their applicants. Each of these companies have been successful with SkillPool, confirming that the software fulfills its purpose as a comprehensive and global resume screening platform.

Chris: What feedback or testimonials have you received from recruiters and businesses that have used SkillPool, and how has their feedback influenced the platform’s ongoing development and improvements?

Soumil: Since SkillPool’s release, we have been dedicated to not only meeting but exceeding the expectations of our users. As a result, the feedback we’ve received from recruiters and businesses who have integrated SkillPool into their recruitment processes has been invaluable in shaping our ongoing development and improvements.

Various recruiters have complimented the accuracy of SkillPool’s AI-driven resume screening, highlighting how it streamlines the initial candidate evaluation process, saving them significant time and effort.

Additionally, our commitment to providing an efficient and user-friendly platform has resonated with recruiting teams. SkillPool’s intuitive design and ease of use have been commended as significant time-savers, allowing recruiters to focus on more strategic and human-centric aspects of their work.

At SkillPool, we’re committed to continuously improving based on feedback from our users. As we move forward, we remain dedicated to delivering an even more powerful and efficient tool for talent acquisition professionals.

Chris: Thanks for being with me, any last words? Where can our readers follow you?

Soumil: Thank you for the opportunity! Artificial Intelligence is a rapidly growing technology, and we believe it is extremely important to integrate it into our current processes. At SkillPool, we do exactly that, improving efficiency and productivity while also enabling Human-AI collaboration. I would definitely recommend everyone interested to check out SkillPool at our website. Alternatively, you can connect with me personally on Linkedin.