An agentic recruitment system, built internally

The challenge

Recruitment is an area where the details matter. How a candidate is assessed, which requirements weigh most, how long data is kept, and who makes the decision. SkyeTec wanted to own these choices, not inherit them from a vendor's assumptions. At the same time, we make our living building AI solutions for others, so we wanted to build one for ourselves first, on a real problem with real consequences.

What we built

A complete recruitment system (ATS) covering the whole journey, from application to signed contract.

Application intake straight from the career pages on skyetec.ai, with job postings managed from the system itself.

AI-assisted candidate assessment. The model reads the CV and application and extracts structured facts, while the system computes a score against the requirement profile. The recruiter gets a candidate card with a summary, observations and a reasoned score, as a starting point, not a verdict.

An interview pipeline from the first meeting to the technical round, with a task library and structured feedback between rounds.

Automatic handover to HR, where the employee and contract are created in Huma when someone is hired.

How we solved it

We built it agentically. The AI does a bounded, verifiable piece of work, reading and structuring information, while the scoring happens in deterministic code. The same CV produces the same score, every time, and we can always explain why.

The system was given its own agent interface, an MCP server, so AI agents can read status, create roles and register candidates in a controlled way, behind proper authentication. It is the same kind of agentic integration we build for clients.

Privacy is built in from the ground up. Deletion deadlines are set on intake and adjusted through the process, and personal data is removed from the text before it is sent to the model. Everything runs on Azure, with data and AI processing inside the EU.

Results

A small team delivered a full system quickly, without compromising on traceability and quality.

A human always makes the final decision. The AI supports, it does not decide.

All data and AI processing stays inside the EU, with privacy and compliance built in from the start.

As important as the tool is the way of working. This is agentic architecture, responsible AI and AI-centric development tried out in-house, and it is the experience we bring into client projects.