Phoenix - Your AI AssistantFix your RAG chatbot: LangChain, OpenAI and vector search
Delivery in
3 days
- Views 16
Amount of days required to complete work for this Offer as set by the freelancer.
Rating of the Offer as calculated from other buyers' reviews.
Average time for the freelancer to first reply on the workstream after purchase or contact on this Offer.
What you get with this Offer
Is your document chatbot giving wrong, vague or made-up answers? I will find out why and fix it.
What I check and fix:
- Chunking and document parsing (tables, PDFs, headings)
- Embedding model and vector search settings
- Retrieval quality: hybrid search, re-ranking and metadata filters
- Prompting and answer grounding with citations
- An evaluation set so you can see accuracy before and after
Works with LangChain, LlamaIndex or custom Python code, and OpenAI, Claude or open-source models.
Why work with me: 7+ years of Python, a 5.0 rating on PeoplePerHour with 40+ reviews, and production AI systems built for healthcare and business automation. Full source code and 7 days of free support are included.
Please message me before ordering with a short description of your project so I can confirm the scope.
What I check and fix:
- Chunking and document parsing (tables, PDFs, headings)
- Embedding model and vector search settings
- Retrieval quality: hybrid search, re-ranking and metadata filters
- Prompting and answer grounding with citations
- An evaluation set so you can see accuracy before and after
Works with LangChain, LlamaIndex or custom Python code, and OpenAI, Claude or open-source models.
Why work with me: 7+ years of Python, a 5.0 rating on PeoplePerHour with 40+ reviews, and production AI systems built for healthcare and business automation. Full source code and 7 days of free support are included.
Please message me before ordering with a short description of your project so I can confirm the scope.
What the Freelancer needs to start the work
1. Access to the chatbot code or repository.
2. Sample documents it uses.
3. Ten or more questions where it currently answers badly.
4. Your vector database and model details.