Objective:
The primary goal of this project is to develop an advanced feature that converts natural language prompts into structured JSON. This feature must not only accurately interpret and translate natural language but also integrate tenant-specific data, such as a list of projects and tasks, to ensure precise extraction and spelling of these entities. This integration is crucial for linking the extracted data with corresponding IDs in our database.
Background:
Our initial experiments with OpenAI's ChatGPT-3.5 Turbo indicated potential but were limited by cost concerns. We seek a solution that is both efficient and cost-effective, tailored to handle specific data inputs unique to each tenant's context.
Tasks to be Completed:
Evaluation and Planning: Assess our current approach and plan a comprehensive strategy to incorporate tenant-specific data, like projects and tasks.
Custom Development: Build a solution that not only converts natural language into JSON but also recognizes and correctly spells tenant-specific terms (e.g., project names, task titles). This may involve creating custom models or algorithms.
Data Integration: Implement a mechanism to dynamically integrate tenant-specific data into the conversion process. This includes fetching relevant data from our database to aid the model in accurately recognizing and referencing these entities.
Optimization for Cost and Performance: Ensure that the solution is cost-effective and does not compromise on performance, considering the additional complexity of handling tenant-specific data.
Integration with Existing Systems: Seamlessly integrate the solution with our existing infrastructure, ensuring it works harmoniously with our database and other system components.
Testing and Validation: Conduct thorough testing, particularly focusing on the accuracy of entity recognition and spelling in the context of tenant-specific data. Address any discrepancies or issues that arise.
Documentation and Training: Provide detailed documentation on the operation and maintenance of the feature, including guidelines for future updates or modifications.
End Goal:
A robust, integrated feature capable of converting natural language prompts into JSON format, with the added capability of accurately incorporating and referencing tenant-specific data such as project and task names. This feature should streamline data entry and linking processes, ensuring high accuracy and efficiency in database operations.
This job is already closed and no longer accepting applicants, sorry.
