01 / The problemDT · 02
The problem
Job searches scatter opportunities across websites. The same role can appear more than once, while keeping track of applications and comparing role requirements becomes repetitive.
← Back to selected work
CASE 02 — SEARCH / ORCHESTRATION / AI
Personal Job Finder
An independent local AI job-search system that connects source discovery, deduplication, persistent tracking and role-fit evaluation.
PythonFastAPISearXNGScrapersPersistent trackingLocal models
Independent build · architecture walkthrough available on requestDiscuss the project
EXECUTION MAPDT / 02
01
DiscoverCombine multiple job-source scrapers with SearXNG search.
02
ConsolidateDeduplicate repeated listings and preserve job state.
03
CoordinateUse a FastAPI orchestrator to organise the search workflow.
04
EvaluateAnalyse CV evidence and job requirements against a candidate profile.
05
RankPrioritise opportunities based on the candidate-job match.
SYSTEM DESIGN / INDEPENDENT BUILD
THE BUILD
Problem → decisions → proof.
A concise view of what was built, why the system is structured this way, and which engineering choices mattered most.
02 / The implementationDT · 02
The implementation
Multiple source scrapers and SearXNG feed a central workflow. Deduplication and persistent tracking organise the data; AI components help plan searches, analyse a CV, interpret requirements and rank opportunities.
03 / The key decisionDT · 02
The key decision
Separate source discovery, durable tracking and model-assisted evaluation. The model helps with interpretation, while the surrounding system owns collection and state.
04 / What this demonstratesDT · 02
What this demonstrates
Multi-source data collection, deduplication, persistence, API orchestration and practical AI-assisted ranking. This is an independent project, not a production hiring service.
DESIGN PRINCIPLEThe useful part is not simply adding an LLM. It is the workflow around it: clean inputs, stable state, clear provenance and a ranking process the user can inspect.
MORE WORKInterested in the