Skip to main content
ProjectsOperational case studyPlatform

Paxi engineering ecosystem

Job Search Automation System

AI-Powered Career Operations Platform · Active

Full-cycle AI-powered job search automation system — RSS ingestion from 10+ job boards, GenAI CV tailoring, cover letter generation, gap analysis reporting, ATS detection, Excel-backed application tracking, and LinkedIn connection targeting. Built and operated as a live daily pipeline.

Delivery classification

AI systems · Automation

Domain tags

platform-case-study · domain-ai · private

Capability tags

ai, ai-native, Python, Claude API, RSS / Atom, openpyxl, DuckDuckGo Search, HTML, PowerShell

Operational environment

  • Replace hours of daily manual job search work with a governed, AI-integrated pipeline — scraping, scoring, tailoring, applying, and tracking applications with full audit history and ATS risk filtering.
  • Modular Python pipeline: RSS ingestion across 10+ job boards, GenAI tailoring engine via Claude API, Excel-backed application tracker, Workday ATS blocklist via URL signal and company-name matching, JD frequency analysis, gap analysis reporting, and daily HTML output surface. LinkedIn connection targeting via DuckDuckGo search with deduplication and personalized outreach generation.

Responsibilities & ownership

  • Replace hours of daily manual job search work with a governed, AI-integrated pipeline — scraping, scoring, tailoring, applying, and tracking applications with full audit history and ATS risk filtering.

Architecture & systems

  • Modular Python pipeline: RSS ingestion across 10+ job boards, GenAI tailoring engine via Claude API, Excel-backed application tracker, Workday ATS blocklist via URL signal and company-name matching, JD frequency analysis, gap analysis reporting, and daily HTML output surface. LinkedIn connection targeting via DuckDuckGo search with deduplication and personalized outreach generation.
  • Representative technology surface: Python, Claude API, RSS / Atom, openpyxl, DuckDuckGo Search, HTML, PowerShell.

Technology stack

  • Python
  • Claude API
  • RSS / Atom
  • openpyxl
  • DuckDuckGo Search
  • HTML
  • PowerShell

Challenges

  • Delivering a governed, maintainable system in a production or production-adjacent enterprise context — balancing scope, security, observability, and delivery cadence.
  • Strategic initiatives often operate under confidentiality constraints that limit public disclosure of architecture detail and quantitative outcomes.

Solutions implemented

  • Replace hours of daily manual job search work with a governed, AI-integrated pipeline — scraping, scoring, tailoring, applying, and tracking applications with full audit history and ATS risk filtering.
  • Modular Python pipeline: RSS ingestion across 10+ job boards, GenAI tailoring engine via Claude API, Excel-backed application tracker, Workday ATS blocklist via URL signal and company-name matching, JD frequency analysis, gap analysis reporting, and daily HTML output surface. LinkedIn connection targeting via DuckDuckGo search with deduplication and personalized outreach generation.

Outcomes

  • Quantitative client outcomes for this initiative are reserved for qualified conversations; the narrative here reflects scoped intent, architecture, and delivery posture as documented in the portfolio source data.

Lessons learned

  • Modular Python pipeline: RSS ingestion across 10+ job boards, GenAI tailoring engine via Claude API, Excel-backed application tracker, Workday ATS blocklist via URL signal and company-name matching, JD frequency analysis, gap analysis reporting, and daily HTML output surface. LinkedIn connection targeting via DuckDuckGo search with deduplication and personalized outreach generation.
  • Operational credibility comes from narrow stack choices, explicit governance hooks, and disciplined extension points — not from surface breadth.

AI-ready metadata

Structured tags and classifications on this page mirror the knowledge object in data/case-studies.ts for future grounding pipelines — no claims beyond the sourced portfolio text.