Heuristics & Confidence Scoring Developer

About Us

The Chamber of Us (TCUS) is a nonprofit building Libelle, an open and transparent volunteer-matching tool. Most résumé screeners are black boxes. Libelle is different: explainable, auditable, and designed for fairness.

We are close to launching our MVP and need help improving how our parser reads résumés, using clear, explainable rules (not black-box machine learning).

Why This Matters

Hundreds of skilled volunteers have already expressed interest in TCUS. To match them with meaningful projects, we need better logic to extract fields (skills, education, experience, location, name) from text résumés. This role helps us get there quickly.

Responsibilities

• Design and implement heuristics to extract key résumé fields (skills, education, experience, location, name).
• Create an explainable confidence-scoring scheme.
• Document rules clearly so non-developers can review and iterate.
• Collaborate asynchronously with other volunteer developers.

Qualifications

• Strong analytical and logical thinking; ability to turn ambiguity into clear rules.
• Comfortable with Python for text parsing (string operations, tokenization, small utilities).
• Experience designing heuristics (beyond regex) and balancing precision/recall.
• Familiarity with fuzzy matching (Levenshtein/rapidfuzz), token overlap (Jaccard), or dictionary-based lookups.
• Familiarity with Google Sheets/Drive APIs (nice to have, not required).

What You’ll Gain

• Real impact: your work will directly enable us to connect volunteers with opportunities.
• Hands-on experience in rule-based NLP and confidence scoring.
• A short, visible contribution to an MVP that launches this fall.
• A chance to be part of a growing, global, mission-driven network.

Commitment

This is a short-term role (2–3 week sprint, flexible hours). You’ll work remotely and collaborate asynchronously with our distributed volunteer team.