ML Ops EngineerLocation: Alpharetta, GA who can work onsite from Day 1 (5 days)Duration: Longterm ContractSkills Required:4-8 years’ experience of applied machine learning/ML Ops in BFS / Investment Management industryPhD or MS in Computer Science, Statistics or related fieldExpertise in Machine Learning algorithms and frameworks:Training and tuning pre-trained modelsWorking with structured and unstructured for Fraud modelsDeep proficiency in Python with experience developing production-quality Python modulesStrong domain focus on fine-tuning and enhancing fraud detection modelsDeploying models in AWS production environmentsStrong command on AWS cloud stack with working knowledge of architecture components i.e., SageMaker, Bedrock, Lambda, Lex, CloudWatch, CloudTrail, Redshift ML, DynamoDB, CodeBuild, CodeDeploy, S3, EC2, IAM, AMIsProficient in API development using Fast API, Flask, etc. delivering asynchronous AI inference services and scalable API solutions for AI-powered applications.Good command over statistical principles of data and model quality e.g., PSI, model performance metrics etc.Roles and Responsibilities:Work closely with Onsite Lead, Data scientists, Data Engineers, and QA and client stakeholders.Evaluate input data for various statistical properties i.e., data drift using PSI and other metricsDevelop methods for monitoring data and models and efficient processes for updating or replacing old models with ones trained on new data or with the latest, state-of-the-art, pretrained models availableSkilled in evaluation metrics like precision, recall, F1-score, and AUC-ROC, ensuring high accuracy and precision in classification and regression models for Fraud.Ensure right-fitting of architecture in AWS for the models at hand to optimize model inferencingStrong working command of AWS SageMaker, MLFlow, and CloudWatch is a mustShould have hands on experience with deploying CI/CD Pipelines in AWSAssist with documentation and governance of all ML and NLP pipeline artifactsFind innovative solutions that increase automation and simplify work in AI workflowsRefactor and productionize research code, models and data while maintaining the highest levels of deployment practices including technical design, solution development, systems configuration, test documentation/execution, issue identification and resolution.