LOCATION: Hyderabad, India (Hybrid) - adjust to team preferences
EXPERIENCE: 1 - 3 Years
JOB TYPE: Full time
ROLE SUMMARY:
The Data Scientist - Video Analytics will work on computer vision models for analyzing real-time videos and detecting hygiene,wastage and environmental-related hazards. The role involves working closely with the Lead architect and the interns. This role involves handling work related to modelling and annotation, testing and evaluating the performance, supporting the deployment.
KEY RESPONSIBILITIES:
- Design data models labeling structures for hygiene, wastage, and environmental-safety detection tasks.
- Develop and maintain video annotation workflows and ensure annotation quality.
- Train, fine-tune and evaluate object detection and action recognition models using kitchen video data.
- Optimize the models to meet sub-2-second inference latency targets on edge hardware.
- Supporting the integration of trained models for the edge-inference and orchestration pipeline in production.
- Track and monitor the deployed models for performance, accuracy degradation, and drive retaining activities.
- Collaborate with the Lead architect on the weighted Hygiene Trust Score algorithm.
- Document model performance, limitations, and bias related findings for AI review.
REQUIRED SKILLS:
- 1 - 3 years experience in video analytics (data modelling, annotation, model training and deployment).
- Knowledge of Python and deep learning.
- Experience with PyTorch or TensorFlow.
- Familiarity with model evaluation metric (mAP, precision/recall)
- Understanding data annotation and tools like CVAT, Labelbox.
- Knowledge on object detection tools(YOLO. Detectron2).
PREFERRED SKILLS:
- Comfortable working in food safety, retail or industrial safety computer vision.
- Familiarity with MLOps tools like MLflow or Weights & Biases.
- Basic understanding of action recognition or pose-estimation models.