DEV Community

Rudra Sarker
Rudra Sarker

Posted on

SightlineAI + Qwen: Building an AI Accessibility Assistant for Blind Users

 «An accessibility-focused AI assistant powered by Qwen, FastAPI, and edge-AI-ready architecture to help visually impaired users better understand and navigate their environment.»


🌍 Why I Started Building SightlineAI

More than 2.2 billion people worldwide live with some form of visual impairment, yet many advanced assistive technologies remain:

  • expensive,
  • inaccessible,
  • closed-source,
  • or dependent on constant internet connectivity.

I wanted to explore how modern AI systems could help create a more affordable and open accessibility solution.

That idea became SightlineAI.

Originally, the project started as a low-cost smart glasses concept combining:

  • computer vision,
  • embedded systems,
  • voice interaction,
  • and assistive navigation workflows.

During the Global AI Hackathon Series with Qwen Cloud, I expanded the project into a full AI-powered accessibility MVP using Qwen and Alibaba Cloud Model Studio.


🧠 What SightlineAI Does

SightlineAI is designed to function as an AI accessibility assistant for blind and visually impaired users.

The current hackathon MVP focuses on:

  • environmental understanding
  • obstacle awareness
  • navigation guidance
  • accessibility-focused AI reasoning
  • structured assistive responses

Users can describe a scene or environment and receive contextual guidance powered by Qwen3.7-Max.


✨ Example Workflow

User Input

«“A hallway with a chair blocking the path and a person approaching from the left.”»

AI Response

«“There is an obstacle directly ahead. Move slightly right to avoid the chair while remaining aware of the approaching person from the left side.”»

The goal is to make AI guidance:

  • practical,
  • concise,
  • safety-focused,
  • and useful in real-world situations.

⚙️ Tech Stack

SightlineAI currently includes:

  • Qwen3.7-Max
  • Alibaba Cloud Model Studio
  • FastAPI
  • Python
  • OpenAI-Compatible API Integration
  • Responsive Frontend UI
  • Accessibility-Focused Prompt Engineering
  • Structured JSON AI Responses
  • Edge-AI-Ready Architecture

🏗️ System Architecture

The MVP architecture includes:

  • Frontend Interface

    • scene/environment input
    • response display
    • accessibility-focused interaction
  • FastAPI Backend

    • request validation
    • prompt orchestration
    • response formatting
  • Qwen Integration

    • contextual reasoning
    • environmental guidance
    • structured AI outputs
  • Future Edge AI Layer

    • wearable hardware
    • computer vision
    • OCR
    • real-time sensors
    • offline inference workflows

🔍 Why Qwen?

One of the most exciting parts of this project was integrating Qwen into an accessibility-focused workflow.

Instead of building a generic chatbot, I wanted the AI system to behave more like:

  • an assistive companion,
  • a navigation helper,
  • and a safety-aware accessibility assistant.

Using Alibaba Cloud Model Studio and the OpenAI-compatible Qwen API, I was able to rapidly prototype:

  • contextual reasoning workflows,
  • accessibility prompting,
  • and structured assistive outputs.

🧩 Challenges

Some of the biggest challenges included:

  • Designing meaningful accessibility-focused prompts
  • Creating structured and reliable AI outputs
  • Building clean frontend/backend orchestration
  • Keeping the architecture scalable for future edge-AI integration
  • Balancing simplicity with real-world usefulness

🚀 Future Roadmap

Future development goals include:

  • 📷 Real-time computer vision
  • 🎤 Voice interaction
  • 🧠 Persistent memory systems
  • 🌐 Offline-first AI workflows
  • 👓 Smart wearable integration
  • 📍 Route learning
  • 🛰️ Edge AI deployment
  • 🔊 OCR + text-to-speech pipelines

❤️ Open-Source Accessibility

A major goal of SightlineAI is keeping assistive technology:

  • open,
  • affordable,
  • customizable,
  • and globally accessible.

Many advanced accessibility devices cost thousands of dollars. Open-source AI can help lower that barrier and allow developers, students, NGOs, and researchers to collaborate on impactful accessibility solutions.


📂 GitHub Repository

🔗 GitHub:
https://github.com/rudra496/sightlineai


🧪 Hackathon Build

This project is currently being developed for:

🏆 Global AI Hackathon Series with Qwen Cloud

Track:

  • EdgeAgent

Focus Areas:

  • Accessibility AI
  • Edge AI
  • AI Assistance
  • Contextual Environmental Guidance

📌 Final Thoughts

SightlineAI is still evolving, but this hackathon MVP demonstrated how modern AI systems like Qwen can be integrated into accessibility-focused workflows quickly and effectively.

I believe AI should not only improve productivity — it should also improve:

  • accessibility,
  • independence,
  • and quality of life.

If you are interested in:

  • accessibility technology,
  • assistive AI,
  • edge computing,
  • or open-source AI systems,

I would love your feedback and collaboration.


🏷️ Tags

AI #Accessibility #Qwen #AlibabaCloud #FastAPI #Python #OpenSource #EdgeAI #Hackathon #AssistiveTechnology #MachineLearning #ComputerVision #IoT #Blind #Innovation

Top comments (0)