About Lydo

What is Lydo?

Lydo is a recruiting signal intelligence platform that analyzes LinkedIn posts before you publish them. Using machine learning models trained on real recruiting data, Lydo predicts:

  • Audience Composition - Which professional roles your post will attract
  • Recruiting Risk - Whether your post signals Helpful, Harmless, or Harmful recruiting patterns
  • Narrative Signals - Underlying themes like toxic culture, burnout, elitism, etc.

How It Works

1. Role Composition Model

Multi-label classifier predicting which professional roles (Software Engineer, Product Manager, etc.) will find your post relevant. Shows percentage distribution across all detected roles.

2. Risk Classification Model

Three-class classifier (Helpful/Harmless/Harmful) that evaluates recruiting risk. Combines model predictions with rule-based checks on narrative flags.

3. Narrative Detection

Multi-label classifier identifying underlying themes like toxic_culture, burnout, credibility_overclaim, culture_misalignment, and elitism.

Features

Single Draft Analysis

Analyze one post to see detailed role composition, risk assessment, and narrative signals.

A/B Testing

Compare two draft variations side-by-side with delta calculations for all metrics.

Contrast Mode

Visual scaling option to emphasize differences in probability distributions.

Raw Model Outputs

No LLM wrappers or explanations - just clean, direct model predictions.

Technical Stack

Frontend

  • • Next.js 16 (App Router)
  • • React 19
  • • TypeScript
  • • Tailwind CSS

Backend

  • • FastAPI
  • • scikit-learn models
  • • TF-IDF vectorization
  • • Python 3.10+

Use Cases

  • →Recruiting Teams: Audit job posts and company content before publishing
  • →Hiring Managers: Test different framings of role descriptions
  • →Content Strategists: Understand audience composition for LinkedIn posts
  • →Employer Brand Teams: Identify and mitigate harmful recruiting signals