We are seeking an experienced developer/team to create a real-time conversational AI system, similar to vapi.ai. The system should be capable of handling real-time conversations, leveraging state-of-the-art technologies for text-to-speech, speech-to-text, and natural language processing. The project must be delivered in dockerized microservices, and all services should be RESTful accessible using WebAPI. Below are the detailed requirements for the project:

Key Requirements:

LLM Model Integration:
– Utilize Llama 3 (https://huggingface.co/meta-llama) as the language model.
– The model should be locally hosted.
– Download the model from here: https://huggingface.co/meta-llama/Meta-Llama-3-8B.

Text to Speech (TTS) Conversion:
– Implement TTS using Unreal Speech (https://unrealspeech.com/).
– Include options for background sounds.
– Enable backchanneling (e.g., ‘mhmm’, ‘ya’) to make conversations sound natural.
– Provide settings for voice speed and temperature.

Speech to Text (STT) Conversion:
– Use Deepgram (https://deepgram.com/) for STT.
Features to include:
– Speaker Diarization
– Smart formatting
– Automatic language detection
– Deep search
– Keyword boosting
– Multichannel support
– Callbacks
– Custom Function Calls:

The system should be able to call custom functions configured within the prompt to the LLM.

Twilio Integration:
– Integrate with Twilio for handling inbound and outbound calls.

Audio Recording:
– Enable audio recording with speaker diarization as an option.

Call Summarization:
– Implement functionality to summarize the call.

Additional Requirements:
– Dockerized Microservices:
– Each component of the system should be delivered as a separate dockerized microservice.
– Ensure that all microservices can be deployed and managed independently.

RESTful WebAPI:
– All services should be accessible via RESTful WebAPI.
– Ensure APIs are well-documented and follow standard RESTful practices.

Deliverables:
– A fully functional real-time conversational AI system with the above features.
– Dockerized microservices for each component.
– Comprehensive documentation covering the setup, usage, and maintenance of the system.
– Source code and deployment instructions.
– API documentation for all RESTful WebAPIs.
– User guide for configuring and using the system.

Preferred Qualifications:
– Experience with LLM models, especially Llama 3 or OpenAI models.
– Proficiency in integrating TTS and STT services.
– Familiarity with Twilio API for call management.
– Strong skills in software development, microservices architecture, and RESTful WebAPI.
– Excellent documentation and communication skills.

Project Timeline:
– Expected completion within 8-12 weeks from the start date.

Budget:
– Please provide a detailed quote with a breakdown of costs, delivery timelines and expectations.

How to Apply:
– Read Requirements Completely.
– Serious and Qualified candidates should only apply. All others would be disqualified and it would be a waste of your time.
– Submit your proposal outlining your approach to this project.
– Include relevant experience and examples of similar projects you have completed.
– Provide a timeline and cost estimate for the project.
– We look forward to working with a talented developer/team to bring this project to life!

Posted On: July 08, 2024 08:25 UTC
Category: AI Integration
Skills:Node.js, RESTful API, C#, C++, Python, Natural Language Processing, API, Artificial Intelligence

Country: United States

click to apply