AMAZON SAGE MAKER
Location : pune
Experience : 4 to 9 Years
1. Amazon Sage Maker
Job Description:
Amazon Sage Maker Key Responsibilities :
1. Design, build, test and maintain end to end ML/DL pipeline
using AWS sagemaker for computer vision and NLP problem statements - to empower data scientists to
rapidly iterate on model development
2. Design and implement end-to-end DL pipelines using AWS Sagemaker for - Processing variety of data sources - Performing data pre- or post-processing, feature engineering, and data validation - Training models using experiments and AutoML/Hyperparameter
tuning using AutoPilot - Performing distributing training by model parallelism and data parallelism using
frameworks like Parameter server, Horovod etc., - Build model evaluation workloads on Amazon
SageMaker - Inference model using Elastic inference, Batch Transform - Deploying models as Endpoints/
exposing as APIs using Flask/Django frameworks
3. Design and develop enterprise AWS solutions for
scalability, reliability and performance
4. Be proficient to implement incremental training and MLOps.
5.
Implement monitoring and self-healing mechanisms for self-recovery.
6. Establish ML Lineage tracking using Sagemaker Preferred Technical Experience
: 1. Hands-on experience in building end to end
pipeline using AWS Sagemaker for State of the art(SOTA) models like BERT, Faster R-CNN, ResNet50, Inceptionv3, EfficientNet etc.,
2. Hands-on experience in data pre-post processing, training, evaluation and deployment using AWS sagemaker APIs for Tensorflow, Pytorch etc.,
3. Hands-on experience in processing both structured and unstructured data from S3, databases like RedShift/NoSQL, Amazon
Kinesis Data Streams etc.,
4. Hands-on experience in processing notifications/messages through Amazon
SNS topic and AWS SMS queue
5. Working knowledge on designing orchestration using AWS-Lambda
and Step Functions
6. Experience in User management using IAM and Monitoring using Cloudwatch
7.Exposure to building large-scale AI/ML solutions using AWS APIs like Rekognition, Comprehend, Transcribe, Translate etc.,
8. Experience in creating custom docker images and pre-defined sagemaker images and registering with ECR
9. Hands-on experience in using Sagemaker Built-in algorithms and building custom SOTA models using sagemaker
10. Exposure to using SageMaker Debugger to inspect training parameters and data throughout the training process
11. Good working knowledge on compute instances, spot-instances and auto-scaling mechanisms
12. Experience in implementing batch processing of data through Batch Transforms
13. Proficiency in Python
14. Active engagement with AWS opensource community.
Nice to Have:
1. Knowledge on Sagemaker GroundTruth for Label Data
2. Experience in monitoring status change events in Amazon SageMaker using Amazon EventBridge.
Job Requirements: ,Amazon SageMaker, Artificial Intelligence, Data Analysis
