From job post tooffer accepted.
A focused recruitment platform built for Soundlines — driving, sales, operations, and workshop hires out of Jebel Ali. What follows is every surface, end-to-end: the candidate journey, the operations console, and the AI layer that ties them together.
What the driver sees, from job post to interview invite.
11 candidate-facing features. Every step is resumable, and the status page reflects the candidate's current stage in real time.
Public job posts
The /roles index lists every open Soundlines position — driving, sales, operations, and workshop — anchored to the Jebel Ali base. Each role page shows the requirements, benefits, and application progress before sign-up.
- ·Salary, rotation, and benefits surfaced above the fold
- ·Requirements and mandate listed in plain language
- ·Three-step application progress indicator (Profile · Screening · Interview)
- ·Single primary call-to-action: Begin Application
Email + password authentication
Candidates sign up to track their application; admins sign in to access the operations console. Email/password and Google sign-in are both supported, with persistent sessions so candidates can resume mid-application.
- ·Email and password sign-in and sign-up on a single screen
- ·Google sign-in available alongside email/password
- ·Persistent session so candidates can resume mid-application
- ·Row-Level Security enforces candidate vs. admin data scopes
Three-step application form
Personal details, eligibility (visa, license category, experience), and CV upload — broken into a stepper so candidates don't abandon a single long form.
- ·Step 1 — Personal: name, email, phone, nationality, location
- ·Step 2 — Eligibility: visa status, UAE license category, years of experience
- ·Step 3 — CV upload via drag-and-drop, stored privately and encrypted at rest
- ·Server-side CV parsing kicks off AI screening in the background
AI CV parsing & summary
On submission, the uploaded CV is parsed by AI — key skills, license types, languages, employment history, and any red flags are extracted and stored as structured data on the application.
- ·Structured extraction: employers, roles, durations, licenses, languages
- ·Plain-language AI summary surfaced on the candidate's status page
- ·Red-flag detection (gaps, license mismatches) flagged to admins
- ·Runs automatically — candidates and admins never wait on a manual review queue
Adaptive AI screening questions
Once the CV is parsed, a tailored screening questionnaire is generated covering compliance, license knowledge, experience, aptitude, and logical reasoning. Resumable mid-session.
- ·Question categories: compliance, license, experience, aptitude, logical
- ·Mixed types: yes/no, single-choice, multi-choice, numeric, free-text
- ·Per-question progress bar; previous answers restored on resume
- ·AI scoring produces a 0–100 overall score plus per-category breakdown
On a GCC cross-border haul, your tachograph card fails mid-route. What do you do?
Live AI video interview
Candidates who pass screening sit a live, two-way video interview with an AI recruiter. The conversation is transcribed in real time, scored against a rubric, and the recording is stored privately for admin review.
- ·Webcam-based two-way conversation with an AI recruiter
- ·Live transcript captured as the candidate speaks
- ·AI rubric scoring across communication, role fit, and competency
- ·Recording stored privately and surfaced in the admin Video tab
Live application status
A live status page shows the candidate's current stage on the six-step pipeline, the AI score, and — once selected — the formal offer letter ready to accept or decline.
- ·Six-stage tracker: Applied → AI Screening → Video Interview → AI Decision → Offer → Decision
- ·Auto-refresh every 8 seconds — no manual reload
- ·Selected candidates see the formal offer with Accept / Decline actions
- ·Rejected candidates see the reason and stage of decision
Notifications inbox
Every status change writes a notification visible from a bell in the header — unread dot, click-to-jump-to-application, 'Mark all read' action.
- ·Real-time bell with unread count (auto-refreshes every 30 seconds)
- ·Per-item unread dot + tinted row until read
- ·Click a notification to jump straight to the candidate detail
- ·Mark all read without leaving the panel
Candidate console — My Applications
A signed-in candidate sees every application they've submitted with submission date, current stage, AI score, and a deep link into the live status page. Auto-refreshes in the background.
- ·All applications listed with date submitted
- ·Current stage label (e.g. 'AI screening in progress', 'Offer received')
- ·AI score badge per application when available
- ·Background refresh every 15 seconds
Offer review & response
When an offer is sent, the candidate's status page shows the full letter, key terms in a side panel, and Accept / Decline buttons with an optional message. Accepting moves the application to Hired and starts onboarding outreach.
- ·Full offer letter rendered on the candidate's application page
- ·Key terms summary: compensation, start date, location, rotation
- ·Accept or Decline with an optional note to the recruiter
- ·Status updates automatically — hired on accept, rejected with reason on decline
Multi-layer AI rejection system
Candidates are filtered at three independent gates so admins only review applications worth their time — and rejected candidates always see why.
- ·CV-fit gate: AI assesses the parsed CV against the role before screening starts
- ·Consistency gate: AI cross-checks screening answers against the CV
- ·Screening cutoffs: hard disqualifier questions and a 55% minimum score
- ·Every rejection logs the stage and reason, notifies the candidate, and renders a dedicated panel on their status page
Where the right candidate gets surfaced, scored, and offered a role.
6 admin-facing surfaces built around one workflow: surface the right candidates, evaluate them on a consistent rubric, and send formal offers without leaving the console.
Operations dashboard
Admins land on a KPI dashboard: total applications, awaiting AI decision, offers ready, offers sent, hired, and average AI score. A funnel chart and recent activity feed live alongside.
- ·Six KPI tiles refreshed every 20 seconds
- ·Funnel chart per pipeline stage (Recharts)
- ·Recent activity stream with deep links to candidate detail
- ·Sidebar navigation: Dashboard · Pipeline · Analytics
Kanban-style recruitment pipeline
Every active candidate, grouped by stage in a horizontal Kanban — search by name/email/nationality, filter to one stage, sorted by AI score then submission date.
- ·Nine stage columns including Video Pending and Video Done
- ·Stage filter chips with live counts
- ·Search across name, email, nationality
- ·Cards show name, AI score, nationality, license, years experience
Candidate detail — five tabs
Per-candidate workspace: Overview (AI summary + parsed CV), Screening (category breakdown + per-question evaluation), Video, Scorecard (1–5 sliders with decision), and Offer (AI-drafted letter, review, send).
- ·Overview tab: AI summary, application details, parsed CV, red flags
- ·Screening tab: per-category score bars + per-question AI evaluation
- ·Video tab: recording link, transcript, AI analysis
- ·Scorecard tab: communication / experience / compliance / culture sliders + decision
- ·Offer tab: AI-drafted offer letter, editable terms, send to candidate
- •8 years HGV Class 1, primarily Dubai–Riyadh and Dubai–Doha lanes.
- •ADR certified · clean tachograph record across last 3 employers.
- •1 employment gap (18 months, 2020–21) — verify in interview.
Quick status actions
Advance candidates from Video Done to an AI-drafted Offer with a single click, or reject with confirmation. Rejections notify the candidate automatically. Hiring is gated behind the formal Offer workflow.
- ·One-click decision from the candidate detail header
- ·Confirmation prompt before rejecting
- ·Hire via offer — the only path to hired is an accepted offer letter
- ·Automatic notification written to the candidate's inbox
- ·Status badge and quick CV download in the same toolbar
Job offer workflow
Once an admin reviewer marks a candidate as 'advance', the Offer tab unlocks: generate an AI-drafted letter, edit terms (salary, start date, rotation, benefits), send to the candidate. A dedicated Offers admin page lists every draft, sent, accepted, declined, and withdrawn offer in one place.
- ·AI-drafted offer letter using role + review notes (Lovable AI Gateway, no API key needed)
- ·Editable terms: position, location, rotation, monthly compensation, start date, benefits
- ·Send → candidate sees the offer on their status page with Accept / Decline
- ·Admin Offers page: filter by status, search by name/email/role, deep-link to candidate
- ·Withdraw a sent offer at any time before the candidate responds
Analytics — pipeline insights
A dedicated Analytics route with score distribution histogram, nationality and license breakdowns (pie charts), pipeline snapshot count, and an offer funnel with live acceptance rate.
- ·Score distribution bucketed into 0–40 / 41–60 / 61–80 / 81–100
- ·Nationality pie chart from real application data
- ·License-category pie chart for fleet planning
- ·Pipeline snapshot: live count by every stage
- ·Offer funnel: draft / sent / accepted / declined / withdrawn + acceptance rate
Built around a focused role catalogue, not a generic ATS.
Driving, sales, operations, and workshop — all anchored to the Jebel Ali base. Each role has its own screening questions, rubric, and AI scoring. No generic 'tell us about yourself' filler.
The AI parses CVs, asks the right follow-up questions, and scores answers — so the operations team spends their time on the strongest candidates, not paper-sorting. Every hire decision is still made by a human.
Candidates see where they stand at all times — current stage, AI score, the formal offer letter, and the reason if rejected. No 'we'll be in touch' silence.
See it from thecandidate's seat.
Submit a real application — you'll experience the full pipeline from CV upload through AI screening to live status tracking.
Begin application →