Swatisama G.

Experience

• 

  Accenture

  Jan 2014 - Jul 2022

  -Led the development of a fraud-tracking banking app, achieving 40% fraud rate reduction.-Designed and implemented an end-to-end ML pipeline for fraud detection, encompassing feature engineering, model selection, hyperparameter tuning, model training, and model serving.-Built a scalable ETL pipeline with Apache Spark and Google Cloud, reducing data ingestion time by 25% and optimizing model training.-Implemented dbt data quality checks, improving data governance and cutting data errors by 40%.-Led MLOps implementation for fraud detection tracker, resulting in a 15% improvement in model deployment efficiency by optimizing the ML pipeline and leveraging Docker and Kubernetes.-Achieved a 10% increase in fraud detection accuracy and significantly enhanced model quality in production by implementing an ML-model observability framework using Grafana and Prometheus.-Implemented unsupervised anomaly detection techniques, resulting in a 20% reduction in fraudulent losses.-Enhanced model accuracy and reliability by implementing extensive testing, validation, and experiment tracking with MLflow, resulting in a 20% performance boost on unseen data. Show less -Led development and release of ML-based insurance claim processing apps, managing cross-functional teams with Agile and KPI analysis.-Implemented Classification ML models for fraudulent claims with TensorFlow and Scikit-learn, achieving 95% accuracy and 85% recall.-Improved claim prioritization with NLP sentiment analysis, reducing processing times by 30%.-Optimized claim triage processes through A/B testing based on claim complexity and urgency, resulting in a 20% reduction in average claim processing time.-Boosted claim approval rates by 10% by applying the Difference-in-Differences causal inference methodology utilizing a Python-based library, Statsmodels, to assess the impact of policy changes.-Accelerated data processing by 30% through a scalable data pipeline leveraging AWS S3 for storage and Apache Spark for efficient batch processing.-Boosted real-time data ingestion for model serving by 30% by implementing a scalable streaming pipeline with Kafka and Cassandra, ensuring low-latency real-time data processing. Show less -Analyzed product analytics data and implemented Power BI dashboards for an insurance risk assessment app, fostering product improvement and resulting in a 10% revenue increase.-Led the design and implementation of 20+ SQL-based data models for the insurance risk assessment application, resulting in a 20% reduction in processing time for risk assessments.-Improved operational efficiency by 10% and enhanced strategic planning through advanced Python regression models for accurate claim volume and resource forecasting. Show less

  • Application Development Team Lead | Lead ML Data Scientist

    Jun 2020 - Jul 2022

  • Senior Development Analyst | Senior ML Data Scientist

    Jun 2017 - May 2020

  • Application Development Analyst | Data Scientist

    May 2015 - May 2017

  • Application Development Associate

    Jan 2014 - May 2015