Bug: Career recommendations lack validation against job market data, providing misleading guidance

[anshul23102]

Problem Statement

DevPath's career recommendation algorithm does not remove bias or validate recommendations against real job market data. The algorithm may recommend careers that are oversaturated, don't match user skills, or contain outdated information, providing misleading guidance that leads users into wrong career paths.

Root Cause Analysis

Recommendation engine likely uses hard-coded heuristics without regular updates. No feedback loop from actual job market data. No statistical validation against real positions. Bias present for popular vs. niche careers. No user outcome tracking (did recommendation lead to employment?). Algorithm weights not adjusted based on placement success rates.

Solution Overview

1. Integrate real job market data APIs (LinkedIn, Indeed, GitHub Jobs API)
2. Validate recommendations against current job postings and hiring demand
3. Track historical recommendation outcomes (did users get jobs in recommended path?)
4. Implement feedback mechanism to rate recommendation quality
5. Add bias detection and mitigation (ensure diverse career recommendations)
6. Build recommendation confidence scores with uncertainty quantification
7. Add alternative recommendations with reasoning
8. Periodically retrain algorithm on updated job market data

Type of Change

• Bug fix (accuracy/reliability)
• New feature
• Breaking change
• Documentation update

Testing Done

Environment

• Job market data API integration
• ML model validation framework

Manual Testing Steps

Test Case 1: Outdated career recommendations

1. User with web dev skills asks for recommendations
2. Check if recommendations align with current web dev job market

Expected: Recommendations include modern stacks (React, Node.js, etc.)
Actual: Recommendations suggest outdated technologies (Flash, jQuery)

Test Case 2: Oversaturation checking

1. Check recommendations against oversaturated fields
2. Verify algorithm avoids recommending 10,000-applicant jobs

Expected: Algorithm recommends emerging fields with fewer applicants
Actual: Recommends only popular saturated positions

Related Issue

Improves accuracy of career guidance and user outcomes.

Suggested Labels

bug, level:intermediate, gssoc26, ml-accuracy, recommendations, career-guidance, data-validation

[dharini-sharma]

Hii @komalharshita @anshul23102 I would love to take this up for GSSoC'26. I have experience with ML validation and can implement the feedback mechanisms and bias detection to ensure the recommendation engine aligns with real job market data. Please assign this to me. Thank you!