Salman Mubashir

Summary

AI & Full-Stack Engineer with two years of shipping production GenAI systems: source-grounded RAG answer engines, fine-tuned NLP models, and automation pipelines, delivered end to end from data and model to API and web application.

Experience

1. CTO & Lead Developer

   Al-Tahqeeq·Mar 2026 — Present·Remote

   • Architected and built AlTahqeeq end to end, a source-grounded AI answer engine for authenticated Islamic Q&A that cites every claim to its original source.
   • Designed and developed the responsive web application and a citation-linked knowledge base of hadiths, validated at a sub-3% error rate with zero fabricated citations.

   • Generative AI
   • Knowledge Base
   • Web Application
   • Full-Stack

   Read the case study →

2. Full-Stack Engineer

   Paragon·Jul 2026 — Present·Remote

   • Developed XtremeAdventures and companion web applications, deploying dynamic registration engines, interactive leaderboards/mapping, and WhatsApp API payment verification to optimize user experience and conversion rates.

   • Web Applications
   • WhatsApp API
   • Payments
   • Maps
   • Full-Stack

3. Full-Stack Engineer

   Vyka Diamonds·Mar 2026 — May 2026·Remote

   • Engineered a responsive Shopify storefront using custom Liquid, JavaScript, and HTML/CSS, optimizing Core Web Vitals and asset delivery to boost mobile page load speeds and user retention.
   • Integrated secure payment gateways, third-party REST APIs, and automated inventory synchronization pipelines to streamline store operations and backend data flows.

   • Shopify
   • Liquid
   • JavaScript
   • HTML/CSS
   • REST APIs
   • Payments
   • Core Web Vitals

4. AI Developer

   Entropic·Oct 2024 — Feb 2026·Islamabad, Pakistan

   • Designed, deployed, and fine-tuned machine learning models for NLP, classification, and generative AI, optimizing foundation models with domain-specific data for production use cases.
   • Engineered data pipelines for preprocessing, cleaning, and transforming large datasets for model training and inference.
   • Built and deployed RAG pipelines using embeddings, vector search, and LLMs for document intelligence and automated knowledge retrieval.
   • Developed FastAPI/Flask APIs integrating AI models into web applications and containerized services on Azure/GCP with Docker.

   • NLP
   • Fine-tuning
   • Generative AI
   • RAG Pipelines
   • Embeddings
   • Vector Search
   • FastAPI
   • Flask
   • Docker
   • Azure
   • GCP
   • Data Pipelines

Projects

1. GluonAI

   Apr 2025 — Jul 2025

   GenAI backend that generates context-aware, on-brand email templates with tiered subscriptions.

   • Developed FastAPI backend leveraging GenAI to create context-aware email templates using brand data, reducing template creation time by more than 70%.
   • Designed RESTful API architecture integrating PostgreSQL, Supabase Auth, and Stripe for tiered subscription management.
   • Implemented scalable microservices architecture with Docker containerization.

   • FastAPI
   • Python
   • Supabase
   • PostgreSQL
   • Docker
   • Stripe

   Read the case study →Live

2. Tenderz Scraper

   Jun 2025 — Oct 2025

   Scalable scraping and AI document-processing pipeline for government tender data.

   • Engineered scalable web scraping infrastructure using Selenium to extract real-time tender data from 40+ government platforms, processing 1,000+ tender documents daily.
   • Built AI-powered document processing pipeline to parse and summarize unstructured PDF/DOCX files into structured JSON/CSV outputs.
   • Implemented fault-tolerant backend with automated error handling, duplicate detection, and chunk-based processing for large datasets.

   • Pipelines
   • Selenium WebDriver
   • Azure Virtual Environment
   • OpenAI API
   • Data Processing

   Read the case study →

3. Retrieval-Augmented Generation (RAG) System for Document QA

   Oct 2025 — Dec 2025

   Local-first RAG pipeline for grounded PDF question answering.

   • Built end-to-end RAG pipeline for intelligent PDF-based Q&A using local LLM deployment, achieving 88% answer relevance score.
   • Implemented document chunking, sentence-transformer embeddings, and cosine similarity retrieval with vector store.
   • Optimized prompt engineering for context-aware responses and designed binary storage architecture for efficient embedding persistence.

   • Local LLMs
   • Prompt Engineering
   • Pandas
   • NumPy

   Read the case study →

4. VisionMap

   Jan 2025 — Mar 2025

   AI-powered video tour generator driven from a 3D Earth interface.

   • Built AI-powered video tour generator with 3D Earth interface.
   • Developed Python backend orchestrating LLM for location context retrieval and RunwayML API for generative video synthesis.
   • Implemented API layer connecting Unity frontend with AI services for real-time tour generation.

   • OpenAI API
   • RunwayML API
   • Unity
   • Python
   • HTML

   Read the case study →

5. Epilepsy Seizures Detection in Neo-natal EEG using Explainable AI

   Final Year Project·Sep 2023 — Jun 2024

   Deep learning seizure detector for neonatal EEG with explainable predictions.

   • Built a deep learning model for neonatal EEG seizure detection with robust signal preprocessing and integrated Explainable AI (XAI) for interpretable, clinically reliable predictions, achieving 82% accuracy with 85% sensitivity.

   • OpenBCI Ganglion
   • Linux
   • Python
   • Google AI

   Read the case study →

Skills

Education

• BSc Electrical Engineering (Computer Engineering Major)

  Sep 2020 — Jun 2024

  FAST NUCES (National University of Computer and Emerging Sciences) · Islamabad, Pakistan

  CGPA 3.5

Honors & Awards

• Gold Medalist, Spring 2023 Semester

  Jun 2023

  FAST NUCES, Islamabad

  Medal for the highest GPA score.

• Gold Medalist, Fall 2022 Semester

  Jan 2023

  FAST NUCES, Islamabad

  Medal for the highest GPA score.

Languages

• English · Fluent