JobHive BD
A next-gen engineering recruitment platform connecting top tech talent with verified employers — featuring in-browser WebRTC video interview rooms (<80ms latency), AI-driven ATS resume parsing, transparent compensation benchmarks, and real-time candidate pipelines.
- >Low-latency native WebRTC video interview rooms with synchronized candidate evaluation rubrics
- >AI-powered ATS resume parsing, automated skill benchmark scoring, and circular matching engine
- >NestJS modular backend architecture with WebSocket signaling, PostgreSQL transactions, and Prisma ORM
Project Vision & Meaning
Technical hiring is bottlenecked by fragmented tools: recruiters manage applications across emails, conduct screens on third-party meeting links without synchronized rubrics, and rely on arbitrary resume filtering. Candidates suffer opaque compensation and slow feedback loops. JobHive BD unifies the entire technical hiring pipeline into a single ecosystem — from verified circular exploration and AI ATS scoring to in-browser WebRTC technical screening rooms and recruiter candidate pipelines.
Real-World Problems Solved
Engineered an end-to-end recruitment platform with dual surfaces for tech talent and hiring teams. Built an interactive circular catalog with verified compensation ranges and real-time category filtering. Integrated low-latency WebRTC technical interview rooms with live screen sharing, chat, and synchronized interviewer scorecards, eliminating external meeting tools. Developed an automated candidate pipeline that shortens hiring cycles from weeks to days.
Challenges & Solutions
Low-Latency In-Browser WebRTC Technical Interview Infrastructure
Problem Context:Conducting live technical screens requires synchronized video/audio, real-time screen sharing, and interactive coding/evaluation without dropped packets or high latency, especially across varied network connections in Bangladesh and globally.
Implemented custom WebRTC peer connections with STUN/TURN traversal and Socket.IO signaling through NestJS. Engineered adaptive bitrate streaming and sub-80ms media transmission with synchronized interviewer evaluation rubrics, ensuring smooth multi-stream video and screen share during technical evaluations.
AI ATS Resume Parsing and Automated Candidate Skill Benchmarking
Problem Context:Evaluating high volumes of applicant resumes manually results in recruiter fatigue and inconsistent scoring. Converting diverse resume formats (PDF, DOCX) into normalized skill profiles with verified compensation matching required deterministic data structuring.
Built an asynchronous ATS Resume DropZone pipeline that extracts structured metadata, skills, and experience vectors. Implemented an automated scoring algorithm that benchmarks candidate capabilities directly against circular requirements, generating match percentages and instant interview recommendations.
Monorepo Architecture & State Synchronization Across Hiring Pipelines
Problem Context:Coordinating complex hiring workflows — job circular creation, applicant status transitions (Applied → Screened → Interview Scheduled → Offer), and real-time chat — across candidate and employer portals risked stale UI states and race conditions.
Structured a scalable monorepo separating Next.js frontend applications and NestJS modular microservices. Utilized TanStack Query with optimistic cache mutations for instant UI updates, paired with PostgreSQL database transactions in Prisma to maintain strict state machine transitions across candidate interview stages.