ARYAN PAHARIBACKEND & APPLIED AIBack to projects
Job Send Crawl dashboard showing service and provider readiness
SELECTED ENGINEERING / Active open-source developmentAryan Pahari

Job Send Crawl

A review-first job outreach system using graph and vector matching, contextual drafts and a self-hosted cockpit.

Overview · Implementation · Evidence
01THE MOTIVATION

The problem

Job outreach takes time, but sending generic messages is not the goal. The useful challenge is connecting actual engineering projects to relevant companies and keeping a human in charge of the final communication.

STATUS / Active open-source development

02UNDER THE HOOD

The implementation

After the workflow proved useful in my own search, I rebuilt it as an open-source system other developers can run and inspect. The current implementation parses a résumé, analyzes GitHub projects, imports job/company data, builds a Neo4j knowledge graph, indexes projects and jobs in Qdrant, ranks project-company matches, generates contextual drafts, and provides a review cockpit for reviewing runs, matches and drafts before mail processing. It remains under active development, with a documented dry-run workflow.

01

01 / PERSONAL

Started as a private workflow to reduce repetitive job outreach.

02

02 / SYSTEM

Added explainable graph + vector + LLM matching instead of generic templating.

03

03 / OPEN SOURCE

Turned it into a self-hostable developer tool with a UI, API, dry-run mail flow and documented setup.

FastAPINeo4jQdrantRedisGroq / OpenAI / GeminiDocker
03ACTUAL ARTIFACTS

Inside the build

Screens from the application, with the workflow and context for each view.

04THE EVIDENCE

Read the source

Explore the code, run the application where available, or watch the original demonstration.