AR.ABDUL REHMAN
RATTU

APPLIED AI / PRODUCT CASE STUDIES

AI case studies.
Real workflows.

Explore six case studies spanning recruitment, voice commerce, digital legacy, access control, material assessment, and road inspection. Choose a project below to see its interface and engineering approach.

01 / PRIVATE PRODUCT / RECRUITMENT INTELLIGENCE

HireIQ Talent Intelligence

A recruitment automation platform for turning high-volume resume screening into a fast, reviewable decision workflow.

HireIQ combines bulk resume uploads, job-aware semantic analysis, and a transparent scoring model to produce ranked candidate profiles. The interface keeps the recruiter in control: each result can show skill verdicts, career timeline signals, and the rationale behind an accept or decline recommendation.

Node.jsExpress.jsGemini 2.5 FlashMulterHTML / CSS / JavaScriptLocal JSON state
Open the product demo
Days → under 10 minScreening objective
Role-aware reviewInteraction
Ranked shortlistOutput

SYSTEM HIGHLIGHTS

  • Bulk resume intake with in-memory file handling
  • Multi-dimensional candidate scoring against a role
  • Transparent accept / decline reasoning for review

INTERFACE WALKTHROUGH

04 SCREENS
HireIQ recruitment intelligence landing screen with ranked candidate scores
01 / INTERFACE

Product positioning and ranked screening view.

Open screen
HireIQ dashboard showing active roles, resumes screened, and recent screenings
02 / INTERFACE

Recruiter workspace overview.

Open screen
HireIQ new screening workflow with role details and job description
03 / INTERFACE

Role definition before screening.

Open screen
HireIQ candidate review showing evidence and match score
04 / INTERFACE

Candidate-level review and rationale.

Open screen
IMPLEMENTATION NOTE

The product demo link was supplied for this portfolio. Private implementation details remain intentionally summarized.

All case studies

02 / PRIVATE PRODUCT / VOICE COMMERCE

Restaurant Agentic AI

An autonomous restaurant platform unifying voice ordering, live order operations, and hands-free analytics.

Restaurant Agentic AI connects a customer-facing ordering experience with an administrative console. Spoken requests can become orders, route-aware ETA calculations support fulfilment, invoices stream from the session, and managers can ask the dashboard questions that become visual and vocal analytics.

PythonFastAPIGemini 2.5 FlashSupabaseWeb Speech APIOSRM / HaversineChart.jsFPDF / PyQRCode
Voice orderingCustomer surface
Live order consoleOperations
Hands-free queriesAnalytics

SYSTEM HIGHLIGHTS

  • Zero-click voice ordering with speech recognition and synthesis
  • ETA calculation with OSRM routing and mathematical fallback
  • Voice-driven analytics translated into charts and spoken insight

INTERFACE WALKTHROUGH

06 SCREENS
Restaurant Agentic AI customer ordering interface with bakery products and voice assistant
01 / INTERFACE

Customer ordering surface.

Open screen
Restaurant Agentic AI menu with categories and product cards
02 / INTERFACE

Menu discovery and product catalogue.

Open screen
Restaurant Agentic AI product detail modal showing an item, price, and ordering action
03 / INTERFACE

Product detail and order action.

Open screen
Restaurant Agentic AI voice assistant handling a bakery order
04 / INTERFACE

Voice assistant ordering flow.

Open screen
Restaurant Agentic AI order dashboard with status, ETA, and actions
05 / INTERFACE

Operations and order control.

Open screen
Restaurant Agentic AI voice analytics dashboard with a sales chart and AI insight
06 / INTERFACE

Hands-free analytics output.

Open screen
IMPLEMENTATION NOTE

This private case study presents the supplied product brief and selected interface frames; proprietary source remains private.

All case studies

03 / PRIVATE PRODUCT / DIGITAL LEGACY

AI Sanctuary

A private memorial and digital legacy platform designed to preserve voice, memories, and meaningful conversation.

AI Sanctuary combines professional voice cloning, conversational reasoning, speech recognition, and semantic memory retrieval. A React and Next.js web experience and a native Swift / SwiftUI iOS application connect to a secure Supabase backend with pgvector, keeping the experience personal while giving memory a durable structure.

ElevenLabsClaudeOpenAI WhisperSupabase / pgvectorReact / Next.jsSwift / SwiftUI
Voice-firstConversation
Semantic recallMemory
Web + iOSPlatforms

SYSTEM HIGHLIGHTS

  • Voice conversations shaped around a person’s authentic tone
  • Semantic memory search for stories and personal details
  • Web and native iOS surfaces around one secure backend
IMPLEMENTATION NOTE

This private case study presents the supplied product brief and selected interface frames; proprietary source remains private.

All case studies

04 / PRIVATE PRODUCT / APPLIED COMPUTER VISION

Automated Parking Access System

A real-hardware access-control pipeline that detects vehicles, reads plates, checks authorization, and drives a physical gate.

This system connects a live RTSP camera feed to a custom-trained YOLOv8 detector, Arabic and Roman-numeral plate OCR, an authorization database, and an Arduino-controlled gate motor. A Flask dashboard gives operators camera visibility, filtered access logs, and vehicle management without requiring a hosted server.

YOLOv8Arabic / Roman OCRFlaskSQLiteRTSPArduino Uno R3L298N H-bridge
99.2% reportedVehicle detection
96% mAP50 reportedPlate pipeline
Under 2 sec reportedLookup
1,970 SAR reportedBuild cost

SYSTEM HIGHLIGHTS

  • Vehicle detection and plate reading on live camera input
  • Authorization lookup against up to 1,000 registered vehicles
  • Physical gate actuation with an operator dashboard
IMPLEMENTATION NOTE

Metrics on this page are reproduced from the supplied project brief and are not independently audited here.

All case studies

05 / MOBILE / COMPUTER VISION

VisionAlytiX

An AI-assisted mobile workflow for assessing asbestos risk in construction materials.

The application uses image analysis to surface material-risk indications and suggest relevant laboratory tests. Its interface brings assessment, results, safety recommendations, and PDF reporting into one mobile workflow.

  • Mobile application
  • Computer vision
  • Risk assessment
  • PDF reporting

THE WORKFLOW

  1. Assess a material

    Begin an assessment from the home screen and provide a material photo for analysis.

  2. Review risk indications

    Read the displayed risk score, assessment details, and safety recommendations.

  3. Plan the next step

    Review suggested laboratory tests and keep the assessment in a PDF report.

MOBILE EXPERIENCE

03 SCREENS
VisionAlytiX mobile home screen with assessment, photo upload, results, recommendations, and reports
01 / INTERFACE

Assessment home

One place to begin an assessment and revisit its results.

Open screen
VisionAlytiX risk analysis screen with a sample risk score, material details, and safety recommendations
02 / INTERFACE

Risk analysis

Risk indications presented alongside material details and guidance.

Open screen
VisionAlytiX PDF report screen with the analyzed material image and assessment summary
03 / INTERFACE

Assessment report

The analyzed image and assessment summary carried into a report.

Open screen
PROJECT VISUALS

Selected interface screens from the project portfolio.

All case studies

06 / UAV / INFRASTRUCTURE

RoadEye

Drone footage translated into pothole detections and geographic context for road inspection.

Aroma AeroTracking — RoadEye applies YOLOv5 to UAV video, identifies road-surface defects, and maps detections with GPS coordinates. A Flask inference backend and Next.js frontend connect video analysis with an interactive inspection view.

  • YOLOv5
  • Flask
  • Next.js
  • GPS geolocation

THE WORKFLOW

  1. Inspect drone footage

    Use UAV road footage as the visual input for pothole analysis.

  2. Locate road defects

    Detect potholes with YOLOv5 and pair the detections with GPS coordinates.

  3. Review mapped findings

    Move from the analyzed video and coordinate table to a geographic view for inspection.

ANALYSIS & MAPPING

02 SCREENS
RoadEye analysis dashboard with processed drone video and a table of pothole coordinates
01 / INTERFACE

Video analysis

Processed road footage and detected pothole coordinates in one workspace.

Open screen
RoadEye geographic map view showing a detected location on a road
02 / INTERFACE

Geographic context

A mapped location gives the inspection findings a geographic reference.

Open screen
PROJECT VISUALS

Selected interface screens from the project portfolio.

All case studies

HAVE AN IDEA WORTH BUILDING?

Let's make it work.

abdul.rehman@team.rapidetechnologies.com