CV - Ravijit’s Homepage

Work experience

Nov 2023 - Present: Data Engineer @ PowerCo

Jan 2022 - Oct 2023: Data Engineer @ Doodle

Researched and Implemented ZSTD compression method in data lake over snappy saving 35% in storage cost, 30% in data scan cost for Athena. All resulting in 20% faster queries
Created a Data Model for millions of daily records ingestion from kafka events for frontend activity on doodle’s website. This powers user churn and conversion analysis and feature activations
Created a scalable solution for ingestion of user activity from kafka topics to database using kafka connect and airflow which empowered in simplifying the dataflow and helped tracking user’s lifetime value
Researched & Created a solution to track and reduce costs of Athena and S3 by solving small file problem for more effective processing of data as well improve the compression of data in parquet files
Implemented Apache Iceberg to support updation and deletion for GDPR compliance in data lake (S3) using Spark (EMR)
Developed custom airflow operator for specific repeated use-cases like mirroring data from athena to redshift among others
Implemented dbt core to enable data democratisation and create accessible data docs for the entire organisation to leverage and support on data-driven culture

August 2021 - Dec 2021: Data Engineer @ Dresslife

Created data pipelines from kafka for MAB for choosing a model from multiple ML models based on its performance in an online fashion based on conversion rate and average order value which helped improve revenue uplift from 3% to 4.3%
Created multiple variants of SR-GNN (Session based GNN) and successfully deployed them to EKS on AWS and helped delivered 3% revenue uplift for fashion retail in sales and non sale situations

July 2019 - July 2021: Data Engineer @ MAQ Software

Developed almost real time Power BI dashboards for Exec team to look at KPIs for various domains using apache spark and azure data lake
Create Models that served multiple use cases and KPI predictions for business to better understand the data and help them make informative decisions
Setup alerts and data quality alerts for various data flows to monitor SLAs and incident management from the ground up using azure application insights
Work on MLFlow and concept drift (based on wasserstein distance) for ML-Ops for retraining alerts and triggers. Use Memory profiler and statistical profiler to log impact on real time services on the retraining model. Extended MLOps to include Sequential A/B testing to test model performance before exposing it to all users.

Jan 2019 - Jun 2019: Intern @ MAQ Software

CV