Abdul Rafey

Abdul Rafey

Frontend Director & AI Developer | Building Scalable, Investor-Ready Platforms

Based in: Lahore, PakistanMost recently: Associate Software Engineer, TechloyceExperience: 1 year

Hire me

Full Stack Developer and AI Engineer with a track record of leading frontend architecture and building real-time AI agents. From directing the development of investor-ready platforms at TourPin to fine-tuning speech models for agricultural impact, I specialize in bridging the gap between complex backend systems and intuitive, high-performance user interfaces.

Experience

Associate Software Engineer

TechloyceLahore, PakistanworkJun 2026 – Sep 2026

Summary

Website: https://www.techloyce.com/ | https://www.subscriptionflow.com/

What I did

  • Migrated core reporting modules from Laravel (PHP) to Next.js, boosting performance ~30%.
  • Ran unit, feature, and integration testing across critical modules and key integrations (Shopify, Avalara, ShipStation, DocuSign), cutting bugs ~20%.
  • Built an AI-agent-powered reporting module on an existing MCP network to enable real-time client data access.
  • Handled direct client requirements and delivery, improving turnaround on requests.

Results

  • Boosting performance ~30%.
  • Cutting bugs ~20%.
LaravelPHPNext.jsShopifyAvalaraShipStationDocuSignMCP network

Frontend Director

TourPinDominican Republic (Remote)workSep 2025 – May 2026

Summary

Website: https://tourpin.net

What I did

  • Built an investor-ready web platform used in product demos with a responsive UI in React, TailwindCSS, and a scalable component architecture.
  • Implemented secure multi-user access with authentication, role-based access control, and backend-managed sessions.
  • Built an interactive Leaflet map with RESTful APIs and full CRUD functionality, improving geospatial user interaction and frontend maintainability.
ReactNext.jsSupabaseTailwindCSSVercelGitLeafletRESTful APIs

AI Developer Intern

AxonBuildLahore, Pakistan (Remote)internshipJun 2025 – Jul 2025

Summary

Website: https://axonbuild.com

What I did

  • Built real-time AI voice agents using LiveKit, FastAPI, and OpenAI APIs for low-latency conversational interfaces.
  • Implemented a RAG system with Qdrant and Ollama for context-aware responses, integrated into a React frontend.
  • Created a reusable RAG + voice agent boilerplate with modular API architecture and a conversion-focused landing page.
ReactPythonFastAPITailwindCSSLiveKitOpenAIOllamaQdrantChromaDBSupabaseRenderVercelGit

Skills

Technical

JavaScriptJavaScript
PythonPython
Next.jsNext.js
ReactReact
GitGit
Retrieval-Augmented Generation (RAG)
Tailwind CSSTailwind CSS
Unit & Feature Testing
TypeScriptTypeScript
SupabaseSupabase
Qdrant
Node.jsNode.js
FastAPIFastAPI
Natural Language Processing (NLP)
Whisper
Postman
PostgreSQLPostgreSQL
AWS S3
AWS Lambda
SQLSQL
C++C++
AWS EC2
ChromaDB
LaravelLaravel
DockerDocker
PHPPHP

Projects

GrowTech | Agriculture Tech Final Year Project

Collaboratoracademic
Collaborated with GrowTech Pakistan to build an AI-powered WhatsApp assistant delivering agricultural guidance via Urdu voice messages, using Meta WhatsApp Business API and NLP. Fine-tuned OpenAI Whisper on 15+ hours of self-transcribed Urdu/Punjabi speech data, reducing WER from 41.29% to 19.90%. Built a RAG pipeline with Gemma to ground responses in an agriculture knowledge base and generate spoken Urdu replies via TTS, integrated with weather and location insights.
PythonWhisperPyTorchGemmaRAGWhatsApp Business API (Meta)NLPTTSGit
Visit the site

SowSmart | Agriculture Tech

Team Leadacademic
Led a 4-member team to build a full-stack smart agriculture platform (React, Node.js) with ML-based crop recommendation and disease detection. Built a chatbot API for treatment guidance and deployed models on AWS EC2 with S3 storage and Lambda pipelines for automated image inference.
ReactNode.jsTailwindCSSPostgreSQLAWS EC2AWS S3AWS LambdaPythonGitRenderVercel
View the code

Intelligent RAG-based Voice Assistant | Generative AI

Developerpersonal
Built a document-aware voice assistant using RAG with OpenAI, Qdrant, and ChromaDB, including PDF ingestion, chunking, and embedding pipelines. Developed real-time voice interaction using LiveKit with a React frontend, evaluating latency, cost, and scalability trade-offs.
ReactFastAPITailwindCSSLiveKitQdrantChromaDBSupabaseRenderVercelGitOpenAI
View the code

Emotion Detection | Machine Learning

Developerpersonal
Built a real-time emotion recognition system using OpenCV and PyTorch, with live video inference via WebSockets and MJPEG streaming for reduced bandwidth.
ReactFlaskOpenCVSocket.IOPyTorchTailwindCSSGit
View the code

Education

BS Computer Science

Lahore University Of Management Sciences (LUMS)Computer ScienceSep 2022 – May 2026

Coursework

Software EngineeringAdvanced ProgrammingOperating SystemsDatabasesArtificial IntelligenceData ScienceDeep LearningNetwork SecurityData Structures & AlgorithmsNetwork-Centric Computing
Hire AbdulReach out about a role, a contract or a conversation.For recruiters

Abdul's twin is AI, it can make mistakes.

More people on Kanvis

Browse more peopleCreate your own Kanvis