Job Description

Educational Background: B.Tech/M.Tech, Computer Science, PhD (optional) Requirements: Required Skills: 2 years of minimum working experience as a DevOps engineer. 1 year of minimum experience working closely with an ML team. Experience in creating, deploying, and maintaining centralized KubeFlow infrastructure on top of one or multiple kubernetes clusters Proficiency in creating CI/CD pipelines for microservice based architectures using Jenkins Proficiency in python. The candidate should be able to write production grade code in python. Proficiency in Git, docker and docker-compose Experience with kubernetes. The candidate should be comfortable with kubectl and helm. Experience working with tools in AWS ecosystem - particularly with Infrastructure as Code (IaC), CloudFormation, IAM, API Gateway, Lambda, Load Balancers, dynamodb, RDS, ECR, ECS and EKS. Desired Skills: AWS certified developer/solution architect Experience in workflow orchestrations tools like Apache Airflow, Prefect, MetaFlow, Luigi etc. Prior experience/familiarity with machine learning frameworks e.g., PyTorch, TensorFlow, ONNX etc. Experience/Familiarity with model serving in ML and working with frameworks like TensorFlow Serving, TorchServe, KFServing, Seldon, BentoML etc. Experience working with computer vision technologies is a bonus Responsibilities: Work closely with the ML team to plan, build, maintain, and improve an end-to-end MLOps platform on top of KubeFlow for research, model training, logging and model serving Work closely with the ML team, integration team(s) and the cloud administrators to deploy and integrate ML services into a wide range of products Build complex container-based workflows that include multiple data and model components for machine learning applications Designing and implementing CI/CD pipelines with git, Jenkins, and AWS for ML research-based projects Continuously improve latency, concurrency, horizontal scaling, and overall API performance for deployed applications by introducing new tools/technologies crafted for ML Understanding and analyzing the current development and deployment specs for the ML team and propose scopes of improvement and solutions to improve the same