Adrian Ezra founded JamieAi to make the recruitment business more time efficient and cost effective. His start-up, which has only been in operation since September 2017, matches available and qualified candidates to jobs offered by employers.
Through word of mouth and blog posts he has amassed a pool of 9,000 data science professionals. When a job is posted, JamieAi finds the candidate from that pool, then asks them if they are interested in that position. Based on a number of factors, JamieAi decides if they are good match. If the candidate is interested and is a good match, the hiring company gets to see the CV. If not, the candidate’s CV never gets released.
"Every time we send a CV, permission has been expressly given."
Privacy is a top concern at the 16-month-old company. Ezra said, “The candidate takes comfort in the fact that their identity is protected and they have anonymity. As a result, we are fully GDPR compliant because every time we send a CV, permission has been expressly given to us.”
At the moment, there is a lot of human intervention in the matching process with Ezra and his team constantly asking questions of the candidates and the companies to find out their preferences. “We, right now, spend a lot of time with the clients to help them understand our community better. Because we understand our community well, we can help them with the job descriptions, through the best use of the words or the phrases, to get the best possible responses from our community.”
This process of working with the clients is especially important in the data field which is renown for nebulous job descriptions and job titles. When companies use ‘data science’ as a synonym for ‘business intelligence analyst,’ it causes confusion among candidates and makes the post a lot harder to fill.
"We live and breathe data and data jobs."
However, the matching algorithm running in parallel and learning from the human manual process, in order for the matching decisions to be increasingly automated over time. “The matching process is definitely supervised with the aim that we will use less and less human overlay. Once the technology begins to do better than me, for example, then I will switch it over to the technology,” said Ezra.
The founder has worked in the field of executive search since 2000 and so has trained the matching algorithm with his experience of recruitment. He set up the tech company because during those years, he noticed inefficiencies in the recruitment process and heard negative feedback about the industry and wanted to add “a bit more trust and transparency back into the system.”
"We help companies use the correct words and terminology in job descriptions."
Ezra said that JamieAi is best placed to bring some clarity to the industry. He said: “We live and breathe data and data jobs, so we will help companies to use the correct words, terminology, and attraction tools for our community, so we get to understand what is it that they are actually looking for. We get into a lot of detail around ho they write their job description and making their job description efficient is a product in itself.”
The results are dramatic with candidates twice as likely to accept a job description that has been optimised. “We would get a 15% acceptance rate when it is unfiltered or unoptimised and we would get 30% when it is,” said Ezra. He said that is optimisation of job descriptions could be automated in the future but for the moment he is working on making the matching algorithm as accurate as possible.
But why choose data and analytics as the testing bed sector for the algorithm? Ezra said it is because data professionals are needed in every industry and it is probably the hardest area to find talent. “It is probably the hardest industry to crack. If I start with the hardest and I get it right, then the nest industry should be a little bit easier”.