Job Requirements
Role: AI Engineer (Computer Vision)
Role Summary
We are seeking a skilled
AI Engineer with strong experience in
computer vision, machine learning, and data processing pipelines to support the development and optimization of video-based passenger detection systems. The role involves working with large-scale
video datasets, annotation pipelines, model training, performance evaluation, and continuous model improvement.
The AI Engineer will collaborate with architects, data scientists, and annotation teams to
build scalable AI solutions, automate workflows, and enhance model accuracy and reliability across different environmental conditions.
Key Responsibilities
- Data Analysis & Dataset Preparation
- Analyze large volumes of video data collected from pilot installations.
- Categorize datasets based on:
- Seasonal variations (summer, winter, fall, spring)
- Lighting conditions (daylight, dim light, nighttime)
- Weather conditions (rain, snow, sunny)
- Identify challenging passenger interaction scenarios such as:
- Multiple passengers entering/exiting simultaneously
- Passengers carrying large objects
- Scenarios causing door operation issues
- Generate statistical insights and reports to support model improvements.
- Data Annotation Support
- Work with annotation teams to ensure high-quality labeled datasets using the V7 annotation platform.
- Support annotation tasks including:
- Bounding box creation
- Segmentation masks
- Object tracking
- Conduct annotation quality reviews and validate labeling accuracy.
- Develop automated or semi-automated annotation pipelines using pre-trained models to reduce manual effort.
- Maintain documentation of edge cases and complex scenarios encountered in datasets.
- Model Evaluation & Performance Analysis
- Test computer vision models using new datasets collected from pilot installations.
- Measure and analyze model performance using metrics such as:
- Precision
- Recall
- F1 Score
- Detection accuracy
- Identify and analyze:
- False positives
- False negatives
- System response time issues
- Document model performance variations across different installations and operational conditions.
- Model Development & Optimization
- Support development and training of computer vision models for passenger detection and door operation analysis.
- Work with object detection architectures such as:
- YOLO variants
- RT-DETR
- EfficientDet
- Other modern detection architectures
- Perform model optimization for edge devices to meet real-time processing requirements.
- Implement data augmentation techniques to improve model robustness.
- Data Processing & Training Pipelines
- Prepare training datasets including:
- Failure scenarios
- Edge cases
- Environmental variations
- Complex passenger interactions
- Convert annotated data into standardized formats suitable for training.
- Maintain training and validation datasets.
- MLOps Pipeline Support
- Contribute to development and maintenance of ML pipelines using AWS SageMaker.
- Support:
- Automated model retraining workflows
- Dataset versioning
- Experiment tracking
- Model version control
- Monitor model performance in production environments.
- Data Drift Detection & Monitoring
- Assist in building systems to detect data drift in production environments.
- Monitor changes in:
- Visual feature distributions
- Environmental conditions
- Object characteristics
- Support development of dashboards for drift monitoring and model performance tracking.
- Image Quality Model Testing
- Test and evaluate image quality assessment models.
- Analyze factors impacting video quality such as:
- Camera degradation
- Lighting conditions
- Environmental factors
- Support retraining of models when performance degradation is detected.
- Extended Computer Vision Capabilities
- Assist in developing advanced computer vision features such as semantic segmentation.
- Train models to identify objects near door areas including:
- Bags
- Strollers
- Mobility devices
- Ensure models are optimized for edge deployment environments.
Required Technical Skills
Programming
- Python
- Data processing and automation scripting
Computer Vision
- OpenCV
- Object detection frameworks (YOLO, RT-DETR, EfficientDet)
- Semantic segmentation techniques
Machine Learning
- Scikit-learn
- Model training and evaluation
- Dataset preparation and augmentation
Data Tools
- Voxel51
- Dataset management tools
Annotation Tools
- V7 annotation platform
- Annotation automation tools
MLOps / Cloud
- AWS SageMaker
- ML pipeline automation
- Model versioning and monitoring
Data Visualization
- Power BI
- Kibana dashboards
Work Experience
Experience Requirements
- 3–6 years of experience in AI/ML or computer vision engineering.
- Experience working with large video datasets and annotation workflows.
- Hands-on experience with object detection or segmentation models.
- Experience with ML pipelines or cloud-based ML infrastructure preferred.
Education
- Bachelor's or Master's degree in:
- Computer Science
- Artificial Intelligence
- Data Science
