1. Mitigated retry storms by implementing a centralized retry strategy, reducing API calls per failure from 70+ to 10, lowering system load.

2. Implemented Update Snapshot feature allowing customers to modify configurations, addressing a key pain point for 100k+ customers monthly.

3. Implemented alpha stack for IDDS service, increasing beta reliability and reducing stack onboarding time by 60%, saving developer effort.

1. Implemented code improvements & SQL query optimizations in Snowflake, reducing the product page load time from TP99 of 45 sec to 8 sec

2. Developed batch processor service from scratch to batch entities and make bulk api calls using Spring Boot, AWS and distributed locking; reducing api calls by 99.5% and bringing down throttling from an average daily 20,000 failures to zero

3. Implemented action tracker to track key SLA metrics, achieving 99.99% success actions by pushing 20 million+ daily events to SQS asynchronously from different microservices and inserting them in MongoDB to generate reports using Node.js

1. Developed a distributed state machine framework to create advertising campaigns for Amazon and Walmart based on Saga Transactions by taking sequential actions

2. Upgraded Market Insights dashboard using efficient SQL queries on 100GB+ datasets in Snowflake, reducing page load time by 85% to 3 sec and increasing product adoption by 20%

3. Implemented Wrapper service framework to transform existing asynchronous actions into synchronous framework; reducing latency by 81%

4. Implemented Action Handler microservice in Spring Boot which would read messages from AWS SQS asynchronously and make third party REST API calls to Criteo

1. Leveraged prompt engineering and LLMs while working on different agents to convert user queries into SQL statements

2. Utilized vector embeddings to perform a proof of concept for finding similarities between user queries and historical data, optimizing cache usage and retrieval efficiency.