Artificial intelligence software CV matching (developers needed)

Description

We need a developer, programmer or coder that is a specialist with artificial intelligence. Ai and data science. Please read the full project description in detail before submitting a proposal. Generic proposals will not be accepted. We have our platform created we just need this specific software so no design or front end development needed. We also have a developer that will help with implementation into our product once the AI software is created
### Project Brief: AI-Powered CV Matching System
Imagine you upload millions of CVs into ChatGPT and ask, “Find me a maths teacher with 10+ years of experience and KS3 expertise.” ChatGPT reads through all the CVs and instantly returns the best match. That’s exactly what we want—but on a much larger scale and without paying for OpenAI’s expensive APIs. Our system must be able to index and analyse millions of CVs, understand employer queries in natural language, find and rank the most relevant candidates in seconds, and work with open-source AI models to keep costs low.
### Project Description
We are looking for a skilled developer or team to build the backend for an AI-driven recruitment platform. The system should allow employers to search for candidates using natural language queries, such as “Find a maths teacher with 10+ years of A-Level teaching experience and QTS.” The AI must go beyond simple keyword matching and truly understand the meaning of the query, returning highly relevant CVs—just like ChatGPT would, but without relying on OpenAI. We need a cost-effective solution that uses open-source AI models such as BGE or Sentence-BERT and efficient vector search tools such as FAISS or Qdrant to balance speed, accuracy, and affordability.
### Key Requirements
The system must be capable of extracting skills, qualifications, and experience from CVs, converting CVs into a searchable format using embeddings for fast retrieval, and storing millions of CVs in a scalable and efficient database. It must allow employers to enter natural language queries such as “Find a GCSE Physics teacher with QTS” while understanding job-specific terms like KS3, QTS, STEM, and A-Level. The AI must be able to match and rank the best candidates based on relevance. Processing speed is critical, meaning the system should be able to analyse millions of CVs in seconds rather than minutes. It must use vector search engines such as FAISS or Qdrant for fast retrieval and include preprocessing techniques to reduce server load and improve efficiency.
To keep costs low, the system must use free and open-source AI models instead of expensive APIs, avoiding OpenAI’s costs while still achieving ChatGPT-level accuracy. Scalability should be a core feature, ensuring that the system can grow without significantly increasing costs.
### What We Need from You
Your proposal should clearly outline how you plan to build this system. You must specify which AI models will be used to process queries and CVs, such as BGE, Sentence-BERT, or Cohere Embed. The proposal should explain how CVs will be stored and retrieved using tools like FAISS, Qdrant, or Vespa. You must also demonstrate how the system will accurately interpret complex job-specific queries such as KS3, QTS, and STEM. Speed and efficiency are crucial, so we need a detailed plan on how the system will handle millions of CVs in seconds. The total project cost must remain under £7,000, and the system must be completed within 3 to 4 months. Proposals that include examples of past work or demos of similar AI-powered systems will be prioritised.
### Why This Project Matters
This system will transform hiring by making recruitment faster, smarter, and more cost-effective. Employers will be able to find the best talent instantly, while job seekers will have better chances of being discovered. We are open to creative solutions such as preprocessing CVs in advance to reduce server load, caching frequent queries to improve search times, and hybrid AI approaches that combine keyword filtering with AI-powered matching.
### Submission Guidelines
Generic responses such as “I can do this” or AI generated generic responses will not be considered. To be eligible, proposals must include a detailed breakdown of the AI, database, and search approach. They must also explain how ChatGPT-level accuracy will be achieved without OpenAI, along with a clear strategy for balancing cost, speed, and accuracy. The proposal must provide a cost estimate and a timeline to completion. We are looking for developers who can demonstrate a deep understanding of AI-driven search and recruitment technologies. Let’s build something powerful together.