Sushma R.

About Sushma R.

Senior Data Engineer with 10+ years of experience designing and building production-grade data platforms across Automotive (Autonomous/ADAS), Banking domains. Proven expertise in developing scalable ingestion, cleaning, and transformation pipelines that power machine learning use cases including perception, path planning, demand forecasting, and fraud detection.Experienced in architecting cloud-native data workflows on AWS (S3, EC2, SageMaker) to support both research-driven and enterprise-scale ML environments. Strong background in building orchestration frameworks using Apache Airflow and Prefect to automate batch, simulation-based, and high-volume data processing pipelines.Skilled in distributed computing frameworks such as Spark and Dask to process terabyte-scale sensor datasets and high-volume financial transaction logs. Adept at preparing curated, ML-ready datasets for supervised learning, reinforcement learning, and predictive analytics models.Collaborated closely with ADAS, telematics, robotics, embedded systems, and data science teams to enable advanced driver assistance features, motion planning models, and enterprise risk analytics solutions. Built automated validation, monitoring, and dataset lineage mechanisms to ensure high data integrity and repeatability across continuous ML experimentation cycles.Strong understanding of the end-to-end ML lifecycle including data ingestion, feature engineering, dataset versioning, training pipeline support, and continuous dataset refresh for model retraining.

Experience

1. Sr Data Engineer

   General Motors

   Aug 2023 — Present · Austin, TX, US

   Designed scalable ingestion workflows on AWS using S3 and EC2 to handle high-volume raw sensor data generated from road-testing and simulation environmentsImplemented data cleaning and transformation pipelines to normalize timestamped sensor inputs across multiple modalities for downstream ML training useDeveloped preprocessing and transformation workflows to align timestamped multi-sensor inputs for sensor fusion and perception model training.Enabled distributed processing of terabyte-scale LiDAR point clouds using Spark and Dask clusters.Automated batch ingestion workflows using Apache Airflow with monitoring, validation, and retry mechanisms to improve pipeline reliability.Implemented data quality checks to detect corrupted sensor frames before entering ML training pipelines.

Education

• Loyola Degree College, Old Alwal

  Bachelor's Degree, computers

  2012 — 2015