All posts
HireNPAIStartupFull-StackNepalLangChain

Building HireNP: Nepal's First AI Hiring Platform

How I co-founded and built HireNP from scratch โ€” an AI-powered end-to-end hiring platform for Nepal, using Next.js, Python/FastAPI, LangChain, and PostgreSQL.

Ritesh BastolaMay 28, 20266 min read

In April 2026, I co-founded HireNP โ€” Nepal's first AI-powered hiring platform. Here's the story of why we built it, what it does, and the technical decisions behind it.

The Problem: Hiring in Nepal Is Broken

The Nepali job market is growing fast โ€” but the hiring process hasn't caught up. Companies receive hundreds of resumes per role and screen them manually. Qualified candidates get missed not because they aren't good, but because no one had time to read to page 3. Interview scheduling involves chains of emails and no-shows. Follow-up is inconsistent.

The result: companies waste months filling roles. Candidates wait weeks for responses that never come. We decided to fix this.

What HireNP Does

  • AI resume screening โ€” automatically parses and evaluates resumes against job requirements
  • Candidate ranking โ€” ranks applicants by fit score, not just keywords
  • Automated interview scheduling โ€” candidates self-schedule through smart calendar integration
  • AI interview notetaking โ€” records, transcribes, and summarizes interviews in real time
๐ŸŽฏ Goal: reduce time-to-hire from weeks to days for Nepali businesses.

The Technical Stack

I designed the architecture end-to-end as the technical co-founder:

  • Frontend: Next.js 15 (App Router) + TypeScript + Tailwind CSS โ€” fast, server-rendered, SEO-friendly
  • Backend: Python with FastAPI โ€” async, high-performance REST APIs
  • Database: PostgreSQL with Prisma ORM โ€” relational, type-safe
  • AI Layer: LangChain + custom LLM pipelines for resume parsing, scoring, and notetaking
  • Infrastructure: Docker + AWS โ€” containerized, scalable deployments

The Hardest Part: Resume Parsing at Scale

The single hardest technical challenge was building a resume parser that works reliably across the chaos of real-world resumes โ€” PDFs with tables, images, two-column layouts, and inconsistent date formats. We built a multi-stage LangChain pipeline that normalizes input before scoring, which dramatically improved ranking accuracy.

Product Strategy & User Acquisition

Beyond engineering, I led early user acquisition โ€” talking to Nepali recruiters, HR managers, and job boards to understand real workflow pain points before we wrote a single line of product code. That customer-first approach shaped every feature priority.

Try HireNP

HireNP is live and accepting early users. If you're a company hiring in Nepal, or a developer interested in the problem space, I'd love to hear from you.