Prompt engineering - creating a system of agents for user interviews

AI & Machine Learning, Python — incorrectly specified categories?

Project translated automatically. Log in or register, to view the original

It is necessary to create an AI system based on agents that will conduct interviews with an expert.

During the interview (about 5 stages, such as "topic", "specialization", "niche", etc.), questions need to be created to delve deeper into the specialist's expertise (for example, a dentist).

AI agents must be configured to self-check the relevance of the questions themselves.

Also, all knowledge must be stored in vector (and/or graph) databases.

That is, the incoming data can vary (text, PDF/Word files, URLs to pages), so the data will need to be stored differently (cut differently and establish different intersections).

Different embeddings, retrieval models are needed, and it is important to have pre-filtering in vector databases to search for information faster.

Additionally, after the interview, post-processing of the data is required to search for contradictory information in the data from the person and documents.

Please do not write bots or "LLM freelancers" - this is a waste of time for everyone. A person with experience is also needed.

Proposals 3 Rejected 3 Discussions 2

Please select

Dzhavid Dzhafarov

0 0

Projects -
Rating -
Rating 144

Budget: 30 EUR Deadline: 1 day

Hello! My name is Javid, I work with systems based on LLM agents and vector storage, I have built custom pipelines for knowledge extraction, multi-step interviews, and detecting contradictions between sources (including PDF, web, and manual input).

I understand that the task is not about a "bot," but about building a deep expert model:
– Setting up agents with iterative validation and question refinement
– Splitting context by sources and types (text, documents, URL)
– Integrating different embedding models + prefilter (by type, tag, author, etc.)
– Post-processing: clustering, identifying contradictions, knowledge graph

I use LangChain / LlamaIndex / Weaviate / Neo4j / FAISS / OpenAI / Mistral.
I will assess the volume of the task and propose an architecture after clarifying the interview format and data types.

Rate — $30/hour. If interested, we can discuss the working scheme and launch priorities.

Mykhailo Pavlov

60 0

Budget: 2500 EUR Deadline: 20 days

Good day. I am ready to complete the task.
I have extensive experience in creating and training AI models.
I mainly code in Python and hold 2nd place on the platform.

I will develop an AI system for conducting interviews with an expert. I will set up machine learning. I will implement all necessary functionality. I will configure the logic for self-checking the relevance of questions. All necessary data will be stored in a vector/graph database. I will set up pre-filtering. I will implement post-processing of data after the interview.
I will create a user-friendly interface. I will optimize.
I will conduct testing and provide the necessary documentation.

My portfolio: Freelancehunt
Write to me, we will discuss the details and I will get started on the work.

V'yacheslav Varvarov

1 0

Projects -
Rating -
Rating 414

Budget: 25 EUR Deadline: 2 days

Hello, the project is interesting, I often deal with AI both in writing prompts and in creating my own neural networks from scratch. Please write me a private message, we need to discuss the project details.

Marcin Piskuła

0 0

Projects -
Rating -
Rating 60

Budget: 850 EUR Deadline: 10 days

I propose the creation of an AI system based on agents that will conduct interviews with experts in 5 stages (topic, specialization, niche, etc.). AI agents generate questions in real-time and also independently verify their relevance.

Data from the interviews will be stored in vector and/or graph databases, supporting various formats (text, PDF, URL) and efficient filtering. The system will also provide post-process analysis, identifying contradictions between the expert's answers and the source data.

Technologies: Embedding models (e.g., BERT, GPT), vector and graph databases (Milvus, Neo4j), advanced filtering algorithms.

I have many years of experience in creating advanced AI systems and managing data. I have already worked on similar projects, which allows me to quickly and efficiently implement a solution tailored to your needs. The project can be completed within 10 working days.