uRecruits

Sr. Full Stack Developer · Tech Lead

AI-driven hiring platform automating end-to-end recruitment workflows

2025–2026•Ongoing
Next.jsTypeScriptNode.jsNest.jsPostgreSQLPrismaRedisAWSOpenAIWebRTC
ai
U
urecruits

uRecruits

AI-driven hiring platform automating end-to-end recruitment workflows

Next.jsTypeScriptNode.jsNest.jsPostgreSQL
Explore →

Problem Solved

Replaced fragmented hiring workflows (spreadsheets, emails, separate ATS, scheduling tools) with one AI-powered platform that screens, schedules, interviews, and reports — letting small teams hire at the velocity of large ones.

Overview

  • •Architected and led development of an AI-powered recruitment platform from scratch
  • •Built AI screening agents that auto-score resumes against job descriptions
  • •Implemented async video interview workflows with auto-transcription and AI summaries
  • •Designed multi-tenant role-based access for recruiters, hiring managers, and candidates
  • •Built scheduling, calendar sync, and automated email/SMS candidate communications
  • •Developed analytics dashboard tracking pipeline conversion and time-to-hire
  • •Integrated with major ATS, calendar, and assessment third-party platforms

Key Skills Demonstrated

AI Agent ArchitectureMulti-Tenant SaaS DesignWebRTC Video Interview EngineeringEvent-Driven Background Job PipelinesOpenAI Structured OutputsTenant Data IsolationTech Leadership & System DesignAWS Cloud Architecture

Challenges

  • •Designing a multi-tenant data model with strict tenant isolation
  • •Generating reliable AI-driven candidate scoring without hallucinations
  • •Handling video interview uploads, transcription, and summarization at scale
  • •Synchronizing hiring stages across recruiters, candidates, and external ATS

Solutions

  • •Built row-level security via PostgreSQL policies + tenant context middleware
  • •Used structured JSON outputs with OpenAI to enforce deterministic scoring rubrics
  • •Pipelined video uploads through S3 → Whisper transcription → GPT summaries with BullMQ
  • •Designed an event-driven sync layer with idempotent webhook handlers

Key Achievements

  • Cut average resume screening time from ~15 min/candidate to <30 seconds via AI scoring
  • Shipped async video interview workflow with auto-transcription and 1-paragraph AI summary
  • Designed multi-tenant architecture supporting 100+ companies with full data isolation
  • Led the engineering team and set architecture/code review standards for the project

Tech Stack

Frontend

Next.js 14 (App Router)TypeScriptTailwind CSSRadix UITanStack QueryWebRTC

Backend

Nest.jsPostgreSQLPrisma ORMRedisBullMQAWS S3AWS LambdaOpenAI APIWhisper

Deployment

Deployed on AWS with Lambda, S3, RDS PostgreSQL, ElastiCache Redis, behind CloudFront with full CI/CD

Third-Party Integrations

OpenAI GPT-4Whisper TranscriptionGoogle CalendarMicrosoft 365SendGridTwilioAWS S3