Oct 2026 – now · Casablanca, Morocco
10xJobs
Side projectSide project
Context
Software engineers lose hours a week scanning job boards for the few posts that fit. 10xJobs does that scanning for them and only shows posts that match a profile they publish.
My part
I build 10xJobs end to end as a side project: a multi-user SaaS on top of Apify actors, and a pipeline that collects new posts, extracts their fields into a typed record with an LLM, and matches them to each user’s published profiles. The pipeline is tuned for cost, efficiency and performance, since every scraped post and every match has a price.
Stack
Outcome
Live at 10xjobs.vercel.app.
What I'd do differently
I am building an evaluation set from the pipeline's own output, so the LLM follows one defined reasoning approach when it extracts fields from job descriptions, and its accuracy can be measured.



Projects at 10xJobs
All projects →
AI
10xJobs: job ingestion pipeline
Collection, preprocessing, type-safe LLM extraction and matching: the pipeline that turns raw job posts into matches for each user's published profile.

Deep dive
10xJobs: search optimisation with segmentation
Matching jobs to profiles like a hash table: both are reduced to segment keys, so each job is only compared with the profiles in its bucket instead of every profile.