Ankit Aggarwal

SEMLOPS

• Designed and developed a Load Balancing, fault-tolerant web application using Nginx as a reverse proxy server.
• Designed APIs for dataset upload to S3 with chunking and versioning.
• Created API for Pipeline, Run, Experiment. Kubernetes resource management based on user requests.
• API to deployed inference models on high-availability servers, implemented log storage using horizontal partition for scalability, and set up monitoring using message queues.
• Implemented cluster data backup with versioning and optimized storage by remove duplicating files.
• Developed real-time process monitoring with Web Socket integration and multi-language support applications.
• Configured SSO, enabling role-based access control for applications.

Prod Ops

• Designed image data versioning systems to track changes over time.
• Developed API to split datasets into training and testing sets with user-defined configurations.
• Integrated model drift notification systems using pub/sub mechanisms to proactively alert users about performance degradation.

ML Ops

• Model Comparison: Evaluated machine learning models' performance across diverse datasets to ensure optimal selection.
• Data Labeling: Designed and implemented data labeling systems for creating high-quality training datasets.
• Pipeline Node Creation: Developed modular pipeline nodes to streamline the AI/ML model development process.
• Excel Report Automation: Built automation scripts for efficient and dynamic Excel report generation.

• Developed a Health Care Analytics API using Spring Boot for data analytics and reporting. Handling 100 concurrent users.
• Created Angular components and utilized charts library to display interactive graphs for health care analytics data.
• Optimized performance of the Health Care Analytics API through database query optimization.
  Developed test suites using JUnit and Mock data for reliable software testing.
• Worked on Hive database to calculate HEDIS, Star Measures, and HCC coding for accurate measures by designing efficient queries and optimizing large-scale data processing workflows.
• Implemented Master Data Management (MDM) processes to consolidate, remove duplicate, and ensure accuracy of data received from 10 multiple sources, creating a single source of truth for reliable data storage and access.