Before completing PGP-AIML in Great Learning, I was purely in an Infrastructure background on backup and storage. Currently, I am not carrying any non-technical experience. My total experience would be 16+ and 14+, which was entirely in the backup and storage domain. I was a backup and storage administrator before joining this program in the 2019 batch.
The problem which I faced at my workplace, where I applied Artificial Intelligence or Machine Learning at Work – “Deep learning projects were occupying more space and ram consumption while running the project. There was an Error occurrence on the data annotation part, which impacted the model performance.”
While running the project on the Pod man container, the size of the Keras was too huge to fit. Hence, there was a need to reduce the size of the .H5 file to fit on the Containers. Many manual errors have occurred while performing the data annotation on the text and images. This was impacting on-time delivery.
Tools & techniques that I applied to solve the problem:
Keras Quantization and Distillation had reduced the trainable parameters and the size of the .H5 files. This helped us to build the pod main image to deliver the project. Automated validation on the Annotated samples to reduce the manual errors. This reduced the bias on the model development and improved the performance. In turn, it helped out the project deliverable on time.
The Insights & Solutions recommended via these AI/ML concept applications were – In-depth analysis and validation.
The Impact my recommendations generated at the organization –
Improvement in the project delivery on time
This exercise benefitted me in upskilling myself on day to day basis. Great Learning’s Mentored Learning Sessions, their Program Support, and the Video Content, along with the Learning Materials, helped me throughout this journey in the AIML Program.
