The problem
Hiring a senior freelancer means reading profile after profile to find the few that fit the job. elital takes the job description and returns a short ranked list, with a written reason next to each candidate, so you can check the ranking instead of trusting it.
Under the hood
The pasted text is expanded into several search queries. Each query runs through two searches in PostgreSQL: vector search, which finds profiles that say the same thing in other words (Voyage AI embeddings stored with pgvector), and full-text search, which matches the words themselves.
An LLM scores each result from 0 to 100 and writes the note for each candidate. System tests cover the guest flow, shortlisting and search.
What it shows about our AI work
We haven’t shipped an LLM feature for a client yet. elital is where we have shipped them ourselves: search finds the candidates, and an LLM scores each one and explains the score in plain words. Our AI features Sprint adds one LLM feature like these to an app you already run, for example search over your own records.
