Umesh . Email & Phone Number

India

Bengaluru, Karnataka, India

• Envisioned, Architected and Implemented a generic Azure Data Pipeline and Spark based Model for batch processing of enterprise level complex retail data, resulting in saving $5000 per month.
• The pipeline reduced the effort by 70% of previous architecture for all loads.
• Sourced, analyzed, processed, validated, transformed, aggregated and distributed data from more than 5+ sources, using generic Azure Data Pipeline.
• Configured and developed automated data bricks spark replication system from PostgreSQL to Snowflake for 100 tables, which improved loading speed of data by 75%.
• Developed and maintained reporting tool ensure 100% errors were recorded and reported.
• Used airflow to orchestrate the ETL solution that helped improve conversion rate by 20%.

Bengaluru, Karnataka, India

• Implemented a Predictive Analysis project called ‘Cell Site Degradation (CSD)’ that trained a ML algorithm to collate raw network and weather data and predict the network site degradation, which resulting in 32%.
• Involved in design and implemented a shipment tracking and alerting product called ‘Dice-Platform’ using Scala Spark, Hadoop and Kafka, improving operating efficiency by 80%.
• Built an algorithm to match vendor delivery quotation and shipment tracking and alert deviation with 100% accurate results.
• Built an end-to-end user configurable CICD Pipeline that was adopted by more than 20+ projects in the company.
• Developed monitoring and alerting capabilities to ensure 100% of data pipelines were working..
• Implemented a ‘Log Analyzer’ using Spark Framework, Elasticsearch and Kibana.

Bengaluru Area, India

• Implemented and automated distributed ETL system for smart data access to determine the aggregation of a traffic of the calls over a network in a particular area for 15mins, half an hour, weekly and monthly, which.
• Created alerting system, dashboard and monitored to ensure 100% data was processed and transferred on time.
• Participated in Agile planning 40+ data feature request.
• Identified and fine tuned tombstone issue in Cassandra to speed up the insert and update by 80% for 10M transaction data per day.