AI based School students Assessment System
Developed a ML pipeline backend for a platform for assessing academic and non-academic performance of school students. Used different machine learning methods (Regression, Classification, Social Network Analysis (SNA) etc.) for creating assessment models.
Tech Stack
- Programming languages : Python
- Frameworks:
- ML models for classification and regression : Scikit-Learn
- NLP Preporcessing: Spacy (NLP)
- Social Network Analysis: Networkx
- ML pipeline : Kedro
- Model Management: MLflow
- DevOps:
- Container : Docker
- version control and CI/CD: Gitea
- Project Management: Taiga
AI Camera based Wild Animal detection System
Developed an efficient prototype for solving human - animal conflicts and wild animal attacks near forest areas using Raspberry Pi and RTSP camera and connected the device to Telegram bot for testing purposes.
Tech Stack :
- Development board: Raspberry Pi-4 4 GB variant
- Programming Languages: Python, bash
- Frameworks:
- YOLO v5 tiny (object detection)
- opencv (image processing)
- multi threading (simultaneous processing)
- DevOps:
- Gitea (version control and CI/CD)
- Project Management: Taiga
Deep Learning based Smart Security System for Playschool Monitoring
Utilized multiple RTSP cameras to capture multiple livestreams for monitoring children’s activities to detect and notify authorities of any suspicious behavior.
Tech Stack
- Development Board: Jetson Nano 4 GB
- Programming languages: C++, Python
- Frameworks:
- Deepstream
- Gstreamer
- Pytorch (model creation)
- Onnx runtime (model quantization)
- Yolo v5 (detection model)
- OpenCV (image processing)
- DevOps:
- Gitlab (version control)
- Project Management: Taiga