Applying for the UK Global Talent Visa can be overwhelming. Many applicants spend months gathering evidence and often pay for professional reviews before knowing whether they have a realistic chance of endorsement.

To make that process more accessible, I built Talent Visa Evaluator, an open source AI application that simulates the endorsement assessment process using publicly available criteria from UK endorsing bodies. It is designed to provide an evidence based evaluation that helps applicants understand where they stand before investing significant time and money.

The project is fully open source and available on GitHub:

Repository: https://github.com/itamelions/talent-visa-evaluator

What the Project Does

The application guides users through a structured seven phase assessment, starting with career profiling and ending with a detailed evaluation report. Rather than relying on a simple questionnaire, it evaluates applicants against both Exceptional Talent and Exceptional Promise routes, recommending the route that best matches their experience.

The assessment adapts to different professional fields including digital technology, academia, humanities, arts, architecture, and fashion. Each field has its own question set, scoring criteria, and AI evaluation prompt.

Key Features
Seven phase guided assessment with downloadable PDF reports
Dual route evaluation for Exceptional Talent and Exceptional Promise
Career stage profiling to ensure fair assessment based on experience level
Field specific evaluation logic and scoring
Detailed evidence reports with strengths, gaps, and improvement recommendations
PDF report generation using jsPDF
Privacy focused design with no accounts, no sign up, and no persistent data storage
Extensible architecture that makes it easy to add support for new endorsing bodies and professional fields
Technical Stack

The project is built with modern web technologies including:

Next.js 16 App Router
React 19
TypeScript 5
Tailwind CSS v4
jsPDF
next-themes

The evaluator supports multiple AI providers including Claude Sonnet 4, OpenAI GPT 4o, and DeepSeek. Developers can switch providers through a simple environment variable without changing application code.

Designing Reliable AI Evaluations

One of the most important parts of the project is the prompt engineering.

Every professional field has its own dedicated system prompt that instructs the AI to act as a former endorsement assessor providing balanced, evidence-based evaluations. The prompts are designed to avoid common AI mistakes by teaching the model that awards, patents, media coverage, and PhDs are not universally required. Instead, the evaluator recognizes equivalent forms of evidence, avoids double-penalizing missing achievements, and produces structured reports with clear reasoning and actionable recommendations.

A shared prompt builder automatically generates the evaluation framework for new fields, making the system easy to extend while maintaining consistent assessment quality.

Why I Open Sourced It

The goal of this project is not to replace professional immigration advice. It is to make the first step of the Global Talent Visa journey more transparent and accessible.

By open sourcing the project, I hope developers can contribute support for additional endorsement routes, improve the evaluation methodology, and help professionals better understand what strong evidence looks like before submitting an application.

If you are interested in AI, prompt engineering, Next.js, or developer tooling that solves real-world problems, feel free to explore the repository, open an issue, or contribute to the project.