Build Voice ChatBot to sell over the phone
Budget: $25 – $50 USD
Develop an AI voice chatbot over a phone connection that is capable of gathering answers to about 50 questions from a consumer in order to make a sale over the course of the call. The app should be capable of using best in class speech to text and text to speech applications in addition to other APIs. The programmer should be highly capable and experienced in javascript, node and object oriented programming. There needs to be a React site to provide a real time view of the interaction, e.g., what consumer said, whether that passed validation to be accepted as an answer. As this will be an internal tool, the site is less important than the programmers backend expertise. Currently a postgres database is being used. A prototype has already been built which can serve as a starting point, although some portions may need to be refactored. The programmer will need to be able to read, understand and optimize the existing node project that makes heavy use of Object Oriented design patterns.
Knowledge of Twilio a strong plus as this is service is currently be used to make and maintain the call. Additionally, the current prototype uses serverless functions to gather consumer input and execute the speech to text.
Additional features that need to be buit are: 1) a vector database to improve the speed and accuracy of the application; 2) several api integrations to validate data or to pre-fill answers for confirmation; 3) many small scale optimizations and enhancements. The near term goal is to produce an MVP within 6-8 weeks, with long-term employment potential for the right candidate. Post MVP, we'd be looking to test a websocket architecture against the current serverless approach to identify the best performing solution.
Knowledge of Twilio a strong plus as this is service is currently be used to make and maintain the call. Additionally, the current prototype uses serverless functions to gather consumer input and execute the speech to text.
Additional features that need to be buit are: 1) a vector database to improve the speed and accuracy of the application; 2) several api integrations to validate data or to pre-fill answers for confirmation; 3) many small scale optimizations and enhancements. The near term goal is to produce an MVP within 6-8 weeks, with long-term employment potential for the right candidate. Post MVP, we'd be looking to test a websocket architecture against the current serverless approach to identify the best performing solution.