AIPath

Open-source AI tooling for careers: CV parsing, resume scoring, skill matching and interview question generation.

AIPath is an open-source project by SJA Pathway for the AI/ML Engineering track. It is built in the open by the community — contribute, learn, and build real, portfolio-worthy work.

Live demo and task board: aipath.sjapathway.com

What's in here

• CV parser — turn a PDF/DOCX resume into structured JSON (contact, experience, education, skills).
• Resume scorer — ATS-style scoring with actionable, explainable feedback.
• Skill matcher — compare a job description against a CV and surface matched vs missing skills.
• Interview question generator — role-aware question sets with model answers grounded in the candidate's real experience.

Tech stack

Python, FastAPI, scikit-learn, spaCy/transformers, plus TypeScript clients.

Who can contribute

Open source contribution is free and open to everyone, but this track is aimed at people with 2–3 years of professional experience in AI/ML Engineering, because the work is real and largely self-directed.

Please note: open-source contribution gives you experience, references from your commit history, and portfolio material — but it does not include a certificate or a reference letter. Those come with the paid internship plans, which include mentorship, structured tasks, and live company projects.

Getting started

git clone https://github.com/SJA-Pathway/AIPath.git
cd AIPath
python3 -m venv .venv && source .venv/bin/activate
pip install -e ".[dev]"
pytest                            # run the tests
uvicorn aipath.api:app --reload   # API docs at http://127.0.0.1:8000/docs

Or use it as a library:

from aipath import parse_text, score_resume, match_skills, generate_questions

match_skills(cv_text, job_description).missing_required
score_resume(cv_text, job_description).checks        # each check has points + feedback
generate_questions("ML Engineer", cv_text, job_description)

1. Pick an open task on the task board. Good first issues are a gentle start.
2. Click Claim on GitHub to open an issue titled with the task id, so nobody duplicates your work.
3. Fork, branch (feat/<task-id>-short-name), commit, and open a PR.
4. Mark the task "status": "done" in tasks.json in the same PR.
5. Join the Discord and say hello in the AIPath channel.

Project layout

Path	What it is
aipath/cv_parser.py	CV parser: contact details, sections and skills from text, PDF or DOCX
aipath/resume_scorer.py	Explainable 0–100 score made of named checks
aipath/skill_matcher.py	Required vs nice-to-have skill coverage against a job description
aipath/interview.py	CV-grounded interview questions with answer guides
aipath/skills.py, aipath/data/skills.json	Skill taxonomy and extraction, shared with the site
aipath/api.py	FastAPI service
tasks.json	The task board: source of truth for what to work on
site/	The live site (static, Cloudflare Workers). public/js/engine.js is a browser port of the Python engine

The core package has no required dependencies. Install extras as needed: api (FastAPI), docs (PDF/DOCX reading), dev (everything plus tests).

The live site

site/ serves the demo and task board at aipath.sjapathway.com. The build copies tasks.json and aipath/data/skills.json into the site, so the board always matches the repo.

cd site
npm install
npm run dev      # local preview
npm run deploy   # maintainers only: deploys to Cloudflare

The demo runs a JavaScript port of the engine in the browser. If you change the Python behaviour, update site/public/js/engine.js to match (automated parity tests are task WEB-2).

Contributing

• One PR per logical change; keep it reviewable.
• Write a clear description of why, not just what.
• Match the existing code style; add tests where they make sense.
• Be kind in reviews. Everyone here is learning.

Community & links

Discord	https://discord.gg/qqE5EqXfdy
Website	sjapathway.com
All repos	github.com/SJA-Pathway
Contact	ceo@sjapathway.com

License

MIT — see LICENSE.