Resume & Recommendation AI Development
Budget: ₹750 – ₹1,250 INR
I want to streamline how resumes move through my hiring pipeline. The first component is an AI-driven agent—call it a chatbot, micro-service, or autonomous script—that can accept incoming résumé files (PDF, DOC, DOCX), parse them, and pull out contact details, skills, education, employment history, and any customised fields I decide to track next. Once the data is structured, the same agent needs to push it straight into my existing applicant-management system through its REST API so that no manual upload is required.
On top of that parsing layer, I also need a lightweight recommendation engine. Using the newly structured data, it should surface the most relevant candidates for a given role or, conversely, suggest the best-fit roles for an incoming candidate. I’m open to the underlying stack—spaCy, Hugging Face transformers, LangChain, vector databases, or another approach—so long as the model’s suggestions can be justified and tweaked over time.
Acceptance criteria
• Resume files processed with ≥95 % field-extraction accuracy on a 50-doc test set
• Data automatically posted to the management system with no lost fields
• Recommendation output returned via API or webhook in <3 seconds per query
• Setup scripts, clean code, and a short README explaining how to retrain or fine-tune the models
If you have previous experience building resume-parsing solutions or AI-powered matching tools, I’d like to see a quick demo or repo link when we start.
On top of that parsing layer, I also need a lightweight recommendation engine. Using the newly structured data, it should surface the most relevant candidates for a given role or, conversely, suggest the best-fit roles for an incoming candidate. I’m open to the underlying stack—spaCy, Hugging Face transformers, LangChain, vector databases, or another approach—so long as the model’s suggestions can be justified and tweaked over time.
Acceptance criteria
• Resume files processed with ≥95 % field-extraction accuracy on a 50-doc test set
• Data automatically posted to the management system with no lost fields
• Recommendation output returned via API or webhook in <3 seconds per query
• Setup scripts, clean code, and a short README explaining how to retrain or fine-tune the models
If you have previous experience building resume-parsing solutions or AI-powered matching tools, I’d like to see a quick demo or repo link when we start.