“Agent AI” is the concept of that moment. Large and small developers are rushing to build apps to leap the layers needed to adopt generative AI in a specific context…and investors are the most interesting funds of these. I’m in a hurry.
In one of the latest examples, a startup from Israel called Perfect (a platform that improves how recruiters raise and hire candidates for employment) has raised $23 million in seed funding I did.
The recruitment team writes open job postings to know where to run and uses Perefect as co-pilot when triaging inbound responses. Both work perfectly, but they actually compete with corporate tools like recruiters, LinkedIn, and more.
A complete claim to save recruiters 25 hours a week. Perfect said that in a year since quietly opened for business, it has grown its customer base to 200 businesses from just 20 startups. The list includes Fiverr, Etoro, McCann and Coralogix.
From facial recognition to candidate selection
Perfect was founded by Eylon Etshtein, best known for being the founder of the perhaps controversial face recognition startup AnyVision (pivot, rebranded, recently acquired).
Etshtein said the idea of perfection came directly from his experience with Anyvision. So he took a very practical approach to hiring, assessing candidates directly and seeing how the process didn’t expand immediately.
However, being the founder of an AI facial recognition startup was set to find “needles in the haystack.” Etshtein imagined a platform trained to understand who Anyvision wanted to hire.
When Etshtein leaves his daily role after things get complicated with Anyvision – this is his current interest in “Resilience” technology, startups building government services and hardware, military and defense purposes Before – he does what he knows next.
There are many AI-based HR startups in the market. Etshtein and its investors believe that perfect is different. First and foremost, it built the platform from scratch – no large third-party language models involved – build your own vector dataset and train it with data sourced from third-party providers Masu. Etshtein said it usually “cleans” data from other large recruiting businesses and “cleans” them to reuse.
“When we started off perfectly, ChatGpt wasn’t out,” he said. “There was no architecture that actually built a career trajectory algorithm that predicted your past, your present, and your future,” he said.
It still took three years of stealth to build the perfect platform from scratch, but that pre-chat work turns out not to replace the final rise of large-scale language models I did. “LLM is a big payload and scary,” he said. In recruiter terms, “payload” is converted into approximately 50 records of data that are considered, annotated and ordered to create insights at the center of all candidates.
“We need to use our own annotated data, otherwise we won’t get the exact results we get today,” he added.
Funds are announced for the first time today, but they are coming in two tranches. Perfect received approximately $12 million in equity investments a year ago from Target Global, RTP Global, Pitango and others. Most recently, I received an interest-free, safe memo from Youngsson, a former Samsung president on the board of Hana Coventures, Jules Ventures and Arm, which will be converted to equity in the next round.
“In an industry desperate for true innovation with both victims and candidates of outdated manual workflows or half-baked AI solutions, Perfect utilizes its own dataset and integrates it into industry-specific workflows. It’s about fully transforming how recruiting works and fully transforming how automated it works. In a statement, HanaCoventures partner, Rior Proter, said:
Certainly, adoption, an area of focus, becomes a hot spot for people building applications with AI, and it’s no wonder because of inefficient adoption.
Certain jobs and certain well-known companies can be overwhelmed by applicants, and the process of finding the most relevant candidate in a mix is probably inevitably like “finding needles in the haystack.” It’s something. In an interview.
Another extreme is also common. Recruiters want to see a variety of applicants, but due to the confluence of vision, work, or organizational unpopular factors, it rarely applies to everyone. In addition to this, human armies tria the application, allowing AI developers to understand how they were refined when they were recruited.
It’s not just perfect being in the space. Other companies include LinkedIn (which has several AI tools for recruiters and job hunters), Hibob, Doable, Maki, Mercor (just raised funds at a $2 billion valuation last week), and Tezi , Seekout (downsize last year) – for dozens more.
The next steps for startups include further enhancements to the toolset they provide to recruiters. Perfect also wants to focus on the other side of the coin. It could potentially plan free tools to help candidates better target their work-seeking efforts and add data to the startup for future projects.