CV Agent

Analyses a job vacancy against Ivan's CV, identifies gaps, and generates a tailored CV + cover letter. Two backends: local Ollama (free, private) or Claude API (cloud).

---

Scripts

File	Description
ollama_job_agent.py	Local Ollama (llama3.1:8b) — no API key needed
job_agent.py	Claude API (Anthropic) — requires ANTHROPIC_API_KEY
make_cv.py	Batch mode: reads "норм" jobs from Google Sheets, generates all, updates status

---

Quick start — Ollama (local)

# 1. Install Ollama: https://ollama.com
ollama pull llama3.1:8b

# 2. Run
pip install -r requirements.txt
python ollama_job_agent.py
# paste a job URL or description when prompted

Non-interactive (for scripts):

python ollama_job_agent.py --auto "https://linkedin.com/jobs/view/..."

---

Quick start — Claude API

export ANTHROPIC_API_KEY="sk-ant-..."
python job_agent.py

---

Batch mode (Google Sheets)

Reads jobs marked "норм" in column G of sheets linkedin and linkedin_promoted in the "Job Tracker" spreadsheet, generates applications, updates status to "cv готов".

python make_cv.py

Requires service_account.json in the project folder.

---

Output

Each run creates a folder in ~/Desktop/CV Agent/:

Company — Job Title/
  analysis.txt          # match score, gaps, strengths
  CV_Ivan_Meshalkin.Rmd # tailored CV (R Markdown / LaTeX)
  cover_letter.txt      # cover letter

---

How it works

1. Analysis — scores CV match (0–100), lists CRITICAL / MODERATE / MINOR gaps
2. Gap filling — interactive CLI asks about missing experience (skipped in --auto mode)
3. CV generation — rewrites CV in Markdown/Rmd to emphasise relevance for this specific role
4. Cover letter — max 250 words, opens with a specific hook, cites 2–3 concrete achievements