[FEAT]: Replace Gemini API with Local Sentence Embeddings for Email Status Extraction

[lnovitz]

Is there an existing request for this feature?

• I have searched the existing issues

Is there a pain point in the app driving this feature request?

• We currently rely on external Gemini API calls to extract application status information from email data
• This creates latency and potential rate limiting issues
• External API dependencies introduce risk if the service changes or becomes unavailable

Desired Behavior

• A fully local solution using sentence_transformers for extracting application statuses from emails
• Comparable accuracy to our current Gemini-based approach
• Lower latency and zero per-request costs

Criteria for Completion

• Create a Python function that takes email text as input and returns the extracted application status using SentenceTransformer (starter code you can use - up to you: https://github.com/lnovitz/jobseeker-analytics/blob/f62b6f5a397c314a4fe1706b226fdd07014bfa11/backend/utils/minilm_utils.py)
• Label your own job application dataset in a CSV - for each email, determine the correct status and enter it in the status_label column.
• Use your dataset to see what the new solution accuracy is. Achieve accuracy within 5% of the current Gemini API solution when tested on your own email dataset

Anything else?

No response

Code of Conduct

• I've read the Code of Conduct and understand my responsibilities as a member of the jobba.help Community

[github-actions]

This issue is stale because it has been open 30 days with no activity. Remove stale label or comment or this will be closed in 5 days.