Karim Abdelkader Email & Phone Number

United States

Carmel, In, Us

Baltimore, Maryland, Us

San Francisco, California, Us

- Architected a Self Service Platform using GitHub Actions, Terraform, Jira Automations and Slack that is used to launch Labeling Jobs in AWS SageMaker GroundTruth or Deploy Models to SageMaker Endpoints. Read about it here: https://www.veranahealth.com/blog/2023/02/engineering-at-work-verana-healths-new-self-service-platform-for-data-labeling-to-help-improve-efficiency-and-data-quality-for-clinical-research/.- Built ML Infrastructure and CI/CD Pipelines for NLP and CV Models for Medical Text and Image data. Supported the ML Model Lifecycle through setting up SageMaker Model Registry, Training / Processing Jobs and Endpoints. Standardizing and automating workflows, orchestrated using Airflow.- Developed Attribute Based Access Control Policies for Project Based Data Access, Permission Boundaries for Tool Access and Service Execution Roles in AWS IAM and SSO. Created OIDC connections for Okta app integrations with AWS Cognito to use SSO for Authentication.

Noida, Uttar Pradesh, In

- Developed and Productionalized End to End Fully Automated CI/CD ML Pipelines using multiple Cloud Platforms such as AWS, Azure and DataRobot.- Client experience using AWS to deploy and maintain business forecasting models, using services such as AWS SageMaker, Lambda, EC2, EventBridge, VPC, CloudWatch, S3, IAM, etc.- Client experience using Snowflake, Docker, Terraform (IaC), CI/CD tools, Orchestration tools, Model Registry tools, Model Versioning and Code Versioning tools.- Trained colleagues from multiple teams in AWS and Azure, in addition to creating training documentation and video demos. - Represented the Data and Analytics team as the MLOps SME in client meetings and RFPs.

Maastricht, Limburg, Nl

Bachelor thesis written as part of the internship, titled: “Applying Machine Learning Techniques to Find Important Features in NAFLD Cell Images". - Developed a Data Science Lifecycle from Preprocessing of Cell Image Data to Feature Selection and Multi Class Classification- Applied ML Algorithms such as Random Forest, k-NN and SVM.- Analysed results by component, category and dye, with conclusions made on biological importance and biomarkers at the organelle level- Research was conducted in collaboration with an ongoing PhD study at the department of toxicology