Linguistic Model Converter (Normalized English)

1. Project Background & Philosophy

Unlike traditional grammar-first pedagogy, our platform is built on a teaching philosophy focused on immediate conversational fluency. We target perpetual beginners—non-native speakers who have academic knowledge of English but lack the ability to communicate in professional or travel contexts.

Central to this ecosystem is the Normalized English model, authored by Mario Rodriguez Pena. The goal is to move students beyond rote memorization into a functional, career-advancing command of the language.

2. Project Scope & Technical Challenge

We are seeking a technical vendor to develop a specialized Converter Tool to translate standard linguistic data into the specific model/language format required by the Normalized English framework.

• Approximately 60,000 words must be covered and processed.
• Previous successful iterations of related models were developed using Python.
• Standard Generative AI models (e.g., GPT-4, Claude 3.5) are currently producing error rates upwards of 90% when tasked with this conversion. This project requires a bespoke algorithmic approach rather than a purely prompt-based AI solution.
• Extend the existing platform functionality to support this specific conversion logic with high precision.
  

3. Requirements (Must-Haves)

To ensure the integrity of the Normalized English model, the vendor must meet the following criteria:

• Deep experience in Python-based development is non-negotiable, as it is the foundation of our existing environment.
• Experience in Natural Language Processing (NLP) or building custom converters where high accuracy is paramount.
• The general scope of work and milestones will be strictly determined and overseen by ISB Vietnam.
• Ability to succeed where standard LLMs have failed by implementing custom logic or fine-tuned processing.