Salman Qurban profile picture

Hello, I'm

Salman Qurban

AI & Software Engineer

Reisterstown, MD  ·  U.S. Permanent Resident

My LinkedIn profile My Github profile

Get To Know More

About Me

Profile picture
Experience icon

Experience

AI/ML Engineering
Backend Development

Education icon

Education

B.S. Computer Science
NLP Specialist (Codecademy)

8M+ Pages indexed in production RAG system
85% Reduction in legal document review time
90% Research efficiency improvement

I'm an AI & Software Engineer specializing in production-grade AI systems. At Neural Lines I architected ShamelaGPT — a live RAG system serving queries over 8 million pages of Arabic literature — and built AI Legal Document Management System, an agentic LangGraph pipeline that cut legal document review time by 85% for Hong Kong barristers. I bridge ML research and backend engineering to ship tools that work in the real world.

Arrow icon

Explore My

Experience

Python Engineer

Neural Lines

Feb 2025 – Nov 2025  ·  Remote

  • Engineered an agentic LangGraph pipeline for an AI-powered legal document management system, automating document review and case timeline generation — reducing manual review time by 85% (2–3 weeks to 2–3 days).
  • Built intelligent document processing using AWS Textract and Tesseract OCR to parse and segment 1,000+ page legal bundles into context-specific sub-documents, achieving 90% timeline accuracy improvement with automated dispute detection.
  • Developed a conversational AI assistant for real-time case analysis with smart document mapping, increasing lawyer productivity by 70% by eliminating manual document organization.
  • Architected and deployed ShamelaGPT (shamelagpt.com), a production RAG system over 8 million pages of Arabic Islamic literature using AraBERT and Pinecone, improving research efficiency by 90%.
  • Enabled multilingual semantic search and cross-lingual Q&A across thousands of Islamic scholarly texts via Gemini 2.0 Flash, serving students and researchers globally.
  • Tech: FastAPI, LangChain, LangGraph, LangSmith, AWS (Textract, S3), PostgreSQL, Alembic, AraBERT, Pinecone, Gemini 2.0 Flash.
↓ 85% review time ↑ 90% timeline accuracy ↑ 70% lawyer productivity ↑ 90% research efficiency Live in production

Software Engineering Fellow

HeadStarter

July 2024 – Sept 2024  ·  Remote

  • Shipped 5 AI-powered projects in 7 weeks including a medical symptom-checking chatbot (sentence transformers + Pinecone + Llama), a spaced repetition flashcard app with AI-generated cards, and a pantry tracker with inventory intelligence.
  • Competed in virtual hackathons forming cross-functional teams globally, delivering scoped AI products under tight deadlines.
  • Attended workshops led by engineers from Google, Meta, and Amazon on scalable system design and AI product development.
  • Tech: Next.js, LangChain, Pinecone, Llama, OpenRouter API.

Technical Skills

AI / ML

LangChain LangGraph LangSmith RAG Pipelines AraBERT Pinecone Scikit-Learn NLTK Sentence Transformers

Backend & Cloud

FastAPI PostgreSQL Alembic AWS Textract AWS S3 Docker Express.js Node.js

Languages

Python JavaScript SQL C++

Frontend & Mobile

React Next.js Flutter HTML / CSS

Tools

Git Postman Tesseract OCR Selenium BeautifulSoup
Arrow icon

Browse My Recent

Projects

AI Legal Document Management

Agentic AI Pipeline · Neural Lines

LangGraph-powered system that automates legal document review, generates case timelines, and detects disputed events for a major law firm. Reduced manual review time by 85% (2–3 weeks → 2–3 days), improved timeline accuracy by 90%, and boosted lawyer productivity by 70%.

LangGraph FastAPI AWS Textract PostgreSQL LangSmith Gemini 2.0
ShamelaGPT
● Live

ShamelaGPT

Arabic Literature RAG System · Neural Lines

Production RAG system over 8 million pages of the Shamela digital library. Supports multilingual scholarly queries using AraBERT embeddings and Pinecone vector search, improving research efficiency by 90%.

AraBERT Pinecone LangChain FastAPI Gemini 2.0 Flash
Hotel Recommendation System

Hotel Recommendation System

Final Year Project · Bahria University

Sentiment-aware recommendation engine trained on 10,000+ scraped reviews using Selenium, BeautifulSoup, and NLTK for sentiment extraction and scoring.

Python NLTK Selenium BeautifulSoup
News App

News & Media Apps

Cross-platform Mobile · Flutter

Cross-platform mobile apps built in Flutter with real-time REST API integrations for live news delivery and media streaming.

Flutter REST APIs Dart
Arrow icon

Get in Touch

Contact Me

salmanqurban36@gmail.com

LinkedIn icon

LinkedIn

GitHub icon

GitHub