ML Engineer Internship

  | #Efficient Movement Detection#Computer Vision#Google Cloud Platform#Production System Handling

Description

A full-time internship in the AI Team/ Deployment team at PromiseQ. I worked on Production Deployment of Models and Motion detection sub-team where I got hands-on experience in Google Cloud Platforms and DevOps MLOps. The primary technical skills I learned during this internship are the Deep learning model and API deployment using the Google cloud platform, Docker, and cloud storage. I also implemented and improved the motion detection sub-section for the threat detection API and on a Shaky Camera Detection System. Further, I also gained experience in the DevOps sector by working on testing pipelines and automating deployment through GitLab runners.

Collaborated Projects

Production Model Deployment

I worked on Production Deployment of Models, which consisted of the maintaining Threat detection API and image cluster management through Google Cloud Platform (GCP)

Technologies

Logo 1 Logo 1 Logo 1

Designing and developing a efficient motion detection system

Previously to detect the motion detection of the detected objects in the scene promiseQ used DeepSORT which had the problems of having high false positives (FP) due to bounding box jitter and high latency, By using the concepts of Background image extraction we developed an efficient detection algorithm which reduced FP and improved latency.

Technologies

Logo 1 Logo 1 Logo 1

Designing and developing a efficient shaky Camera detection system

Created a dataset , implemented algorithms and tested efficient shaky camera detection techniques.

Technologies

Logo 1 Logo 1 Logo 1

Automating Testing/Deployment Using MLOps and Gitlab runners.

Automated the latency tests, accuracy tests and deployment to dev, stage and production environments/clusters using Gitlab runners.

Technologies

Logo 1 Logo 1 Logo 1

Key Achievements

  • Production Model Deployment.
  • Designing and developing a efficient motion detection system.
  • Automating Testing/Deployment Using MLOps and Gitlab runners.
  • Designing and developing a efficient shaky Camera detection system.