Senior NLP & Prompt engineer required
Budget: $15 – $25 USD
Who we are looking for:
We need a senior full-stack expert to fix our LLM-integrated chatbot server. The project is almost done and if you are an expert, it can be done within a day.
What project is it?
The project is made up of two parts.
- The LLM server which hosted open-source LLM model.
- The fastapi server which is interacting with LLM server and front-end.
The fastapi server is interacting with LLM server with OpenAI and langchain and generates responses for human responses.
The back-end retrieves results from MongoDBAtlas but from multiple collections and based on the retrieval result, it generates the appropriate answer for the human response.
What issues are there?
The main issue is that the chatbot doesn't work well as we expect.
For example, the chain doesn't generate condensing question as expected and thus don't provide appropriate answer.
What should you do?
First, you should review the LLM chain and refactor it.
Second, you should choose the appropriate language model from Hugging Face for our use case.
This is a quick fix project and if this works well, I can work with you in terms of other aspects like testing, deployment and so on.
I look forward to hearing from you.
Thanks
We need a senior full-stack expert to fix our LLM-integrated chatbot server. The project is almost done and if you are an expert, it can be done within a day.
What project is it?
The project is made up of two parts.
- The LLM server which hosted open-source LLM model.
- The fastapi server which is interacting with LLM server and front-end.
The fastapi server is interacting with LLM server with OpenAI and langchain and generates responses for human responses.
The back-end retrieves results from MongoDBAtlas but from multiple collections and based on the retrieval result, it generates the appropriate answer for the human response.
What issues are there?
The main issue is that the chatbot doesn't work well as we expect.
For example, the chain doesn't generate condensing question as expected and thus don't provide appropriate answer.
What should you do?
First, you should review the LLM chain and refactor it.
Second, you should choose the appropriate language model from Hugging Face for our use case.
This is a quick fix project and if this works well, I can work with you in terms of other aspects like testing, deployment and so on.
I look forward to hearing from you.
Thanks