Pratik Asai

AI/ML Data Platform

• Lead the architecture and implementation of scalable AI/ML Data platforms, expertly utilising SageMaker Studio, MLFlow, Airflow and RAG platform using AWS Bedrock.
• Created E2E NLP to Text AI Agents/NPS customer feedback categorization Agent using Bedrock, Lambda over complex Data including data preprocessing via scala spark
• Fast-tracked AI usage in the company using LLM platform for maximizing adoption of AI Platform

Customer Data Management

• Lead and deployed CDM analytics platform to collect 100+ data points of
  customers using user activity on site, order history, etc. for concise targeting of marketing emails to users.
• Developed customized Trigger email solution for 5+ emails including Order and Ship Confirmation emails
• Developed Email and SMS performance data metrics pipeline, CCPA and GDPR
  compliance pipeline and Loyalty bonus cash appeasement pipeline.
• Worked extensively on AWS cost reduction by optimizing of EMR Spark jobs by
  80%
• Developed Data Quality enforcer and automated data ingestion processes to eliminate manual effort for migration.
• Created Feeds pipeline to partners for sharing Customer/Email/Transaction
  Feeds via SFTP.

Customer Recognition

• Integrated with 3rd party tool like Audience acuity and Alteryx to increase
  identification of customer on site by 40%
• Developed new emails payload to extend recognition emails arsenal like cross
  Abandon Cart/Browse.
• Migration of system to Iceberg hive metastore to enhance RW benchmarks.

Fan360

• Lead and developed Deduplication of customer accounts from 135M to 105M
  across sites using customer's email/FN/LN/Address for concise targeting
• Collected data metrics at user level and shared across Fanatics entities using
  Snowflake
• Enabled cross-sell pipeline between "Commerce" and "Betting & Gaming"
  vertical using above data collection for customer.

• Developed Learning Resource Management System analytics data pipeline to feed the dashboard
• Established event-driven system using Maxwell with in-house infrastructure to read CDC from binary logs of MySQL and writes row updates as JSON to Kafka.