From Recruiter to AI Builder: What Federico D'Alessio Learned Building His Own AI Recruiting Tools
Most recruiters use AI tools. Federico D'Alessio decided to build his own.
On the latest episode of Freedom of Work, the podcast from RemoFirst, host Leah Cottham sat down with Federico, a talent acquisition leader turned AI builder, to talk about why he went from running hiring teams at companies like Idealista and Trovit to building AI products like Mike (an AI voice recruiter) and Talent Signal (a sourcing tool that surfaces candidates actively job-hunting in public communities).
The conversation covers where AI genuinely saves recruiters time, where the hype outruns the reality, and what the recruiter role looks like once the admin work disappears.
Why Talent Acquisition Pulled Him In
Federico has spent 15 years in tech and SaaS, moving from HR generally into talent acquisition specifically. What drew him in, he said, was the mix of business goals, hiring manager needs, and candidate fit all sitting inside one role.
That mix is also what pushed him toward automation. After years of watching recruiters repeat the same administrative tasks thousands of times, he started asking a simple question: if the technology to remove that repetition already existed, why were people still doing it by hand?
The Two Moments That Changed His Thinking
Federico points to two specific moments where AI stopped being a buzzword for him.
The first was ChatGPT, around three years ago. He was working with a new remote hire whose written communication style made it hard for colleagues to understand him in chat. Federico started using ChatGPT to bridge that gap and realized the tool could help build working relationships remotely, not just answer questions.
The second was discovering AI voice platforms. Testing a voice AI that could hold a natural conversation was, in his words, close to unbelievable. That's when he started thinking seriously about where the technology could go in five or ten years, and where it could fit into his own job.
What He Actually Built
Mike, the AI voice recruiter, handles the first-stage screening conversation: location, salary expectations, language fluency. It currently runs in four languages, with the platform built to support more than 16. The goal isn't to replace the recruiter. It's to filter out mismatches early, so a 20 to 30 minute screening call doesn't happen with a candidate who was never going to be the right fit for that specific role.
Talent Signal solves a different problem: finding candidates who are actively looking but not saying so on LinkedIn. Developers in particular tend to signal availability in places like Hacker News threads rather than updating their LinkedIn status. Talent Signal pulls that public information into one place and can draft an AI outreach message once a recruiter pastes in a job description. Federico is clear that both the outreach copy and the candidate profile still need a human review before anything goes out.
Neither tool was cheap or easy to build. Federico is upfront that running AI in a professional environment costs real money and time, and that both products went through many rough early versions before becoming usable.
Where AI Actually Saves Time
Asked where AI is genuinely helping versus where it's overhyped, Federico pointed to two clear areas:
- Sourcing. Boolean search used to be a specialist skill. AI now runs that search and returns results, which changes what "being good at sourcing" even means.
- Admin and transcription. Cleaning up notes, updating the applicant tracking system (ATS), and summarizing calls are tasks AI can now absorb, freeing up two to three hours a day for recruiters to spend on relationship-building and understanding the business.
What AI can't do, in his view, is the human part: reading candidate motivation, building trust with hiring managers, negotiating with stakeholders, and understanding what a business actually needs before matching a person to it.
Keeping the Process Human
Leah raised the bias question directly: if AI is involved earlier in the hiring process, how do you make sure the process stays fair and trustworthy for candidates?
Federico answers that the responsibility sits with the humans training and reviewing the system, not the tool itself. He also raised a distinction worth sitting with: the question isn't only "can we automate this?" It's "should we?" Not every task in a hiring process is a good candidate for automation, even if the technology makes it possible.
The Recruiter of the Future
Federico doesn't think AI replaces recruiters. He thinks it changes what the job is. Less manual execution, more strategy, evaluation, and relationship management. Fewer hours lost to Boolean search and ATS updates, more time spent understanding the business well enough to push back on a hiring manager's decision when the data supports it.
His advice for anyone wanting to start: don't over-engineer it. Pick one painful, repetitive task, automate that first with a simple tool, and improve it over time. He compares his first working versions of Mike and Talent Signal to what they are now: barely recognizable.
Federico's tools are built for the recruiting side of hiring. Once a role is filled, the next challenge for global teams is often compliance: paying and managing that person correctly wherever they're based. That's the problem RemoFirst's Employer of Record (EOR) platform solves, handling local payroll, contracts, and compliance in 185+ countries so hiring teams can focus on finding the right people rather than the paperwork that follows.
