PromptHire

Overview

A job search agent that takes a user's resume and a natural-language prompt (e.g. "senior backend roles in Singapore, fintech preferred") and returns ranked, relevant job listings eliminating manual keyword tuning on job boards.

Technical Challenges

Agentic Workflow with LangChain

The core flow is a LangChain agent with two tools: a resume parser and a Google Search API wrapper. The agent decides when to call each tool, extracts structured skill/experience data from the resume, constructs optimised search queries, and synthesises results all in a single invocation. Prompt design was critical to prevent the agent from hallucinating job URLs rather than using the search tool.

Resume Parsing

PDF resumes are parsed with a text extraction library and then passed to Gemini with a structured output prompt to extract skills, years of experience, job titles, and preferred industries. Output is validated against a Zod schema before being used as agent context.

Search Query Construction

Naively passing resume text to Google Search produces poor results. Instead, the agent synthesises a ranked list of skill keywords and combines them with location/industry filters from the user prompt to construct multiple targeted queries, then deduplicates results by URL and re-ranks by relevance using an LLM scoring pass.