Spaces:
Sleeping
Sleeping
Commit History
-
Fix interview scoring so strong answers score well
husseinelsaadi +1: Claude Opus 4.8Scores were capped at ~50% even for perfect interviews because the final score weighted skills-match (exact string overlap, almost always ~0) at 40%. Rebalance to answer-driven scoring and reduce conservative bias: - interview_api.py / report_generator.py: weight QA answers 85% and skills match 15% (was 40/60); bump category values (Good 0.8->0.85, Medium 0.6->0.65). Mirrored in both so screen + PDF report agree. - interview_engine.py: rewrite the evaluate_answer prompt with a clear rubric, tell LUNA to use the full range and not default to "Medium", and to judge substance not grammar (answers are speech-to-text). Output format unchanged, so parsing is untouched. A flawless interview now scores ~85% (more with matching skills); weak answers still score low. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>c3ec385
-
Accurate LLM CV parsing + realistic, memory-aware interview
husseinelsaadi +1: Claude Opus 4.8CV parsing (resume_parser.py, app.py): - Parse resumes with the Groq LLM (clean name/skills/education/experience JSON), far more accurate than the old keyword/regex heuristics; keep the regex/NER path as a lazy fallback so startup no longer loads the heavy NER model - Auto-parse the uploaded CV on apply and populate the profile (manual form fields still override), so an ATS-friendly CV "just works" Interview realism (interview_engine.py, interview_api.py): - Give follow-ups real memory: pass the persisted interview_log as conversation_history (the frontend never sent it) - Evaluate answers at the real job role + seniority instead of a generic default - Enrich prompts with job skills/seniority/description; LUNA now acknowledges the last answer and asks one focused, contextual follow-up - Fix an empty-experience IndexError that silently dropped the first question Also includes earlier deploy hardening: env-driven SECRET_KEY/FLASK_DEBUG and removal of a duplicate generate_next…c28f81f
-
updated
88df9ba -
check up done
cb47676
-
updated
e83166d -
updated
6720344 -
updated
32acb92 -
back version
a1b807c -
interview logic updated
3e2c190 -
report added
b5d3943
-
updated with enum
49fbb3a -
updated
d23da13 -
updated
6d75d91
-
update
27994de -
audio updated
330157f -
full new update
308d699 -
audio updated
5cee863 -
interview updated
44441db -
interview updated
8e4e001 -
updated
81341b4 -
updated
fd40bd4
-
indendent updated
9f2b0ed -
updated
ca3e053
-
updated
22b00f2
-
Initial commit for Hugging Face Space
2ae57cb