Hi, I’m Mehraj Hossain

Applied AI Engineer focused on building end-to-end AI systems for real-world use, with an emphasis on image generation and computer vision.

About Me

My work is hands-on and engineering-driven. I build AI applications from scratch, integrating models, backend services, and frontends into usable systems. I’m particularly interested in how generative models behave under real usage, and many of my projects are designed to surface consistency and control issues through practical applications.

Selected Projects

Virtual Try-On & Wardrobe System

A production-oriented virtual try-on system focused on clothing extraction, wardrobe persistence, and image consistency across poses and backgrounds. Real-world images are segmented, cleaned, and stored for reuse in outfit generation.

  • Clothing segmentation and extraction from real-world images
  • Persistent, user-specific wardrobe and outfit gallery
  • Prompt-based pose, background, and lighting edits
  • Support for multiple user and clothing models

Multi-Modal AI Assistant

A chat-based assistant that integrates text generation, image generation, image editing, and document-based Q&A into a single interface.

  • Text generation using Google Gemini
  • Image generation and conversational image editing
  • Document Q&A using RAG pipelines
  • Shared conversation state across modalities

Core Skills

AI & Machine Learning

  • Applied generative AI systems
  • Prompt engineering and tool calling (LangChain)
  • Image generation workflows (Stable Diffusion, ComfyUI)
  • Computer vision: segmentation, in-painting, extraction
  • Retrieval-Augmented Generation (RAG)

Engineering & Cloud

  • Python, JavaScript, PHP
  • React, Node.js, FastAPI, HTML/CSS
  • Google Cloud (Cloud Run, Vertex AI, GCE)
  • Docker and serverless deployment

Let’s Connect

I’m interested in applied AI roles focused on real-world systems and practical use of ML models.