Project idea
I want to build Avertaq SiteGuard, an edge AI safety and compliance assistant for construction sites.
The system will process a live camera feed locally and detect:
- workers entering a predefined restricted or hazardous area,
- workers not wearing a safety helmet,
- workers not wearing a high-visibility vest.
When a compliance event is detected, the system will save a timestamped event and a relevant image, rather than continuously uploading the full video stream. It will then generate a simple daily safety and compliance report.
Where it will be used
The project is designed for construction sites, warehouses and other industrial areas where safety monitoring is currently performed manually or through continuous CCTV observation.
The initial prototype will use one camera and one predefined monitoring zone.
Why it is useful
Construction companies need a practical way to improve safety monitoring, document incidents and produce evidence for internal compliance reporting.
Running the inference at the edge offers several advantages:
- faster event detection,
- reduced network and cloud requirements,
- improved privacy because the complete video stream remains local,
- operation in locations with limited or unreliable internet access.
How Voyager Wingman will help
I plan to use Voyager Wingman to describe, generate and refine the computer vision pipeline in natural language.
Wingman will help me configure and debug the Voyager SDK pipeline for:
- person and personal protective equipment detection,
- restricted-zone logic,
- live camera input,
- event generation,
- saving evidence images and metadata.
The goal is to demonstrate how a user with limited previous experience of the Axelera AI platform can move from a clearly defined prompt to a working edge AI application within four weeks.
Four-week MVP scope
To keep the project realistic and buildable, the final prototype will focus on:
1. detecting people in a live or recorded camera feed,
2. detecting helmet and high-visibility vest compliance,
3. detecting entry into one configurable restricted zone,
4. recording timestamped compliance events,
5. producing a simple daily event report,
6. demonstrating the complete pipeline on Axelera AI hardware.
A lightweight dashboard may be included if time permits, but it is not required for the core demonstration.
Why I am the right person to build it
I work in a safety-critical aviation environment and have practical experience with procedures, operational discipline and the importance of reliable compliance checks.
At the same time, I am developing Avertaq, an AI-based platform focused on construction tender document analysis and compliance. This project extends that same mission from pre-project document compliance to real-world site compliance during project execution.
I am building my software and AI skills through practical projects, and I am prepared to document the complete process openly, including prompts, technical decisions, problems, progress updates, code and the final demonstration.
Avertaq SiteGuard is intentionally focused: one camera, a small number of clearly defined detections and a measurable working outcome within four weeks.
