Companies have many tools for tracking technical talent, but a new startup wants to help recruiters by bringing together the worlds of GitHub and LinkedIn to create a database of the most suitable candidates for a specific software development job. role – and it does this by using AI to “infer” capabilities from the code you write.
Prog.ai, as the company is called, allows employers to search for developers based on their technical skills, the libraries they use, or simply their contributions to projects on GitHub.
In the year Founded in 2022 out of San Francisco, Prog.i is the brainchild of CEO Maria Grineva, who previously sold data startup Orb Intelligence to Dunn & Bradstreet in 2020. CTO Fedor Soprunov, formerly a machine learning researcher at Russian tech titan Yandex; And Dmitry Pianov, head of product, has worked in product teams at companies including Yandex and Replika.
While recruiting was initially the company’s main focus, as the inaugural product opened to recruiters in closed beta this week, Grineva sees a wider range of use cases beyond helping companies fill technical roles. This includes fostering developer relationships, such as asking them to join a community or inviting them to contribute to an open source project. Ask their experts for a specific problem; And even developer tool companies to help them install their products.
“This week we’re opening Prog.ai to tech recruiters, and in April we’ll be extending our SaaS offering to developer contacts with Prog.ai to help companies building tools for developers realize their TAM (Total Reachable Market). Learn more about their existing developer community, and reach their target audience,” Grineva said. TechCrunch explained.
To help begin its commercial push, Prog.i announced today that it has raised $1 million in pre-seed funding from angel investors, Brooklyn Bridge Ventures, and angel backers including one of Spotify’s original employees and himself. Former CTO, Andreas Ehn.
So how does Prog.ai go about inferring capabilities from crowdsourced code? Well, first of all, the platform runs GitHub’s “git clone” command, which creates copies of millions of public repositories and branches. Prog.ai then analyzes each Git commit, and examines the code snippet, file path, and commit subject to find out what it’s about.
“For a project, we can see who is the main architect, who develops the back or front end, who focuses on UI / UX, who builds QA and tests and who are the technical writers,” said Grineva.
Prog.ai also pokes holes in git actions like pull requests, including rejections and approvals, comments, and open spaces, which are used to help project contributors “understand” the different roles and levels of involvement.
“We run not only popular open source projects, but also ‘pet’ projects, experiments, forks and training projects from Coursera or Udemy that engineers publish on GitHub,” Grineva added. Together, we’re working on nearly 1 billion projects a year on GitHub to get the right profile of each engineer’s skills.
In code, Prog.ai relies on OpenAI’s GPT, customizing the much-talked-about language model in high-profile open source projects and StackOverflow articles to help it score points on code quality, for example.
Prog.ai users can build lists of top experts in specific fields such as “large language models” or “computer vision” and create a leaderboard of top performers in any field. Or you can provide a list of repositories and rank all contributors by word count.
Effectively, employers and companies can tailor their search based on their desired criteria, including skill areas, programming languages, and years of experience.
But understanding code is only part of Prog.ai’s offering.
A major selling point for employers is affordability. meet with software developers, and for that Prog.ai includes a built-in email distribution engine, powered by sales engagement platform Reply.io.
“Users use our search to create a list of relevant candidates, then they can create a personalized email sequence, mentioning candidates by name, mentioning their projects and explaining why they think the position is a good fit for them,” said Grineva. .
Recruiters also want a more rounded view of a developer’s skills, education, and work history, which they probably won’t get from GitHub. This is where LinkedIn enters the fray, with Prog.ai crunching publicly available data and matching it with individual data from GitHub. And that’s what Grineva calls the platform’s special sauce — by combining data from two widely used platforms, it can build a more nuanced picture of potential candidates.
“I believe that the combination of GitHub and LinkedIn profiles brings a lot of value, since engineers are usually not very good at promoting themselves and often do not even have full LinkedIn profiles,” said Grineva. “Also, people self-report on LinkedIn, which means the data is objective. Implementing a standardized method based on the actual code contributions of all engineers not only eliminates the issue, but also means companies can evaluate candidates more consistently.”
Of course, none of these provide a perfect recruiting pipeline. Bringing together two gargantuan, disparate data sets is no easy task, and there can be plenty of room for error here, raising the possibility of conflating profiles with similar names and histories. And that’s assuming someone has a LinkedIn profile in the first place, which may never happen. But under the hood, Grineva says he has put some steps in place to address some of the potential problems with the lease.
“Matching two large data sets is not an easy task, because the information people provide on GitHub can be small and many engineers choose to remain anonymous on GitHub,” Grineva explained. “We’ve built a proprietary fuzzy-matching system that considers not just names, usernames, and email addresses, but jobs, expertise, interests.”
In addition, Grineva said that they use computer vision to compare profile models on platforms, which, while not foolproof in itself, serves as an additional tool to other verification methods.
As of this writing, Prog.I says it has contact information for about 70% of all profiles in its database, but that means about 30% are missing critical information. In the meantime, Grineva said that although they hope to improve the contact list as it expands, the issues that can be used do not always revolve around access.
“Another important use case is data enrichment,” she said. “Clients can search for a full candidate’s profile by GitHub handle, LinkedIn URL, or contact email – in which case we can only match 70% of the email we have.”
Here’s the big elephant in the room: Isn’t Prog.I facilitating “cold callers” to find developers in bulk?
“There’s a risk, but it’s important to know that recruiters are already trying to call developers, and that this is currently happening with other tools, and some tech recruiters are manually pulling contact information from GitHub,” Grineva said. “That said, employers are currently doing this with bad or limited understanding of developers, which means the distribution is impersonal and often the opportunity is not developer-friendly. As a result, these emails come across as spam.”
For those on the receiving end of a Prog.ai-powered outreach campaign, Grineva says the platform is “fully GDPR compliant,” and developers can ask it to remove or edit their profile, as well as opt out entirely. Email delivery.
Show me the money
It’s still early days for Prog.ai and it’s experimenting with different schemes, but the company is essentially running a SaaS-based subscription model, based on the number of contacts a user reaches. This starts with “Free” up to 100 contacts per month, up to the “Employer” plan, which costs $530 per month for advanced search features and 3,000 contacts. It also offers a custom priced enterprise plan, available upon request.
Not to mention other recruiting solutions that cover everything from LinkedIn’s own Talent Solutions product through Zoominfo, SeekOut, TalentOS and HireEZ. But Greneva says Prog.I’s focus on technical talent and Github’s scan smarts is what sets it apart from the crowd. In turn, this can It means better targeted headhunting efforts, where recruiter and candidate goals are more closely aligned.
“Being an engineer myself, I receive a lot of messages from employers that are not relevant to me, and I look into this problem myself,” said Grineva. “I believe this is primarily a data quality issue: employers don’t have enough information about me to match me with opportunities that interest me. Our goal is to reduce the standard that audio developers receive today. By providing employers with better information, we believe this will be a win-win for both developers and employers.