Case study·2026 · AI
SmartHRFlow — AI Recruitment from Resume to Offer
An AI hiring platform that parses resumes at 98% accuracy, scores every candidate, and runs the whole pipeline from application to offer.
- Client
- Confidential · reference on request
- Sector
- AI
- Timeline
- 12 weeks
- Live
- smarthrflow.com
smarthrflow.com — live hero
(01)
The challenge
SmartHRFlow is an AI-powered recruitment platform that takes the manual grind out of hiring — resumes parsed and scored the moment they arrive, assessments that resist cheating, and a pipeline view that carries every candidate from application to offer. Live at smarthrflow.com.
HR teams were losing their weeks to screening: reading resumes in a dozen formats, re-keying candidate details, chasing applicants over email one at a time, and defending shortlists built on gut feel. The brief was blunt — cut the manual workload dramatically, keep the evaluation fair, and make setting up a new job's hiring flow a minutes-long task rather than a project.
(02)
What Metaxia built
A Next.js and TypeScript front end over a NestJS API, with Supabase as the data layer and the OpenAI API doing the reading.
- Resume parsing at 98% accuracy across formats, extracting skills and experience into structured profiles
- Intelligent skill matching with automated ranking and a bias-aware evaluation flow, each candidate scored against the role
- A stage pipeline — applications, shortlist, HR interview, technical interview, offer, hired — with live counts per stage
- Custom skill assessments: multiple question formats, custom scoring, and anti-manipulation safeguards
- Bulk candidate outreach with reusable templates and delivery/engagement tracking
- Drag-and-drop application form builder, mobile-optimized, with a new job's flow live in about five minutes
(03)
Inside the product
hiring pipeline
Every role gets a live pipeline: candidates arrive parsed and scored, then move through named stages with counts at a glance — no spreadsheet shadowing the process.
- Stages from application to hired — shortlist, HR and technical interviews, offer
- Per-candidate AI score against the role, recalculable on demand
- Top-candidate surfacing above the applicant pool
- Search, filter, sort, and stage controls inline
- Credit-based posting with add-candidate and share actions
(04)
Delivery
The platform shipped to production at smarthrflow.com on Vercel, with the assessment and bulk-email modules following the core parsing pipeline. A v2 interface is already in preview, iterating on the recruiter dashboard without interrupting live hiring.
Timeline: 12 weeks · one accountable team
(05)
What changed
Recruitment steps reduced by 60 percent end to end
98 percent parsing accuracy across resume formats
Screening time cut by roughly 60 percent with AI ranking
A new job's application flow configured in about five minutes
Building something similar?