Unbaised360

• 

Inspiration

This project began with something deeply personal. My brother recently received a surprisingly low performance rating despite consistently delivering twice the amount of work as others on his team. He had been an intern manager, took on extra responsibilities, and had strong outcomes — yet one unresolved bug and a manager’s “standard deduction” cost him his rating. Seeing how subjective and unfair the process felt made me realize how many people, especially quiet high performers or underrepresented employees, experience the same injustice. That moment inspired Unbiased360 — a tool built to make performance evaluations more transparent, fair, and evidence-based.

What It Does

Unbiased360 is an AI Performance Analyzer that evaluates employees using a combination of:

Self-review Manager feedback Peer sentiment Evidence & contribution patterns

It generates a fairness-adjusted performance breakdown across:

Work Impact Problem Solving & Quality Team Contribution

And provides an AI assistant that explains results clearly — without revealing backend formulas.

How I Built It

Lovable was used to build the full UI, screens, CRUD flows, and all rating logic.

Toolhouse API was integrated to power an intelligent performance assistant that answers employee questions about their ratings.

Combined both tools to create a cohesive system where UI, logic, and AI interaction work together.

Challenges I Ran Into

My biggest technical challenge was integrating the Toolhouse API agent into the Lovable interface.

Managing the flow between the AI assistant and the dynamic UI required multiple iterations.

Ensuring the agent understood the rating logic without revealing it was also a delicate balance.

Accomplishments That I’m Proud Of

Designing a complete, fully functional performance review system in a short timeframe

Creating a fairness-driven logic model inspired by real-life inequality

Building an AI assistant that explains ratings in human language

Solving a real problem that affects thousands of employees globally

What I Learned

Through this project, I gained hands-on experience with:

Lovable Chat Mode and its advanced UI-building capabilities

Toolhouse API and agent-based architectures

Prompt engineering for safe and controlled reasoning

Building and modularizing scoring logic for fairness systems

What’s Next for Unbiased360

Enhancing the rating model to improve accuracy and fairness

Adding deeper bias detection for managers and teams

Integrating with HRIS platforms (Workday, BambooHR, Rippling, etc.)

Expanding the AI assistant to provide personalized improvement plans

Built With

• lovable
• toolhouse