May 2022 – Jan 2026 · 3 yrs 8 mos · Seattle, Washington, United States · On-site

• Worked with the Alexa Smart Home team, focused on designing and developing scalable ETL pipelines using SQL, Spark, Python, EMR, Redshift, and other AWS technologies to support Alexa Plus services. This work enhances conversational AI capabilities, particularly for predictions and routines in Smart Home’s LLM initiative.
• Designed scalable microservices and APIs using Python and AWS for Smart Home LLM latency tracking (Time-To-First- Token/ Time-To-Full-Output) with real-time monitoring and alerting pipelines which in turn reduced latency by 35%.
• Led compliance initiatives for all Smart Home data stores which ensured safe, secure and compliant AI systems.
• Designed an internal tool to manage a Redshift Data Warehouse containing multi-petabytes of data, which accounts for 80% of all Smart Home datasets, leading to a 12% increase in efficiency and an 8% reduction in costs.
• Analyzed large-scale data in Data Lake, built end-to-end data pipelines and delivered reporting solutions to provide actionable insights for stakeholders in the Alexa Smart Home, Devices team and Marketing team.

Nov 2019 – Mar 2022 · 2 yrs 4 mos · New York City Metropolitan Area · On-site

• Worked as an AI/ML Engineer on Optical Character Recognition, and automated Signature Verification for financial forms and travel receipts to reduce manual effort by 40%.
• Trained and deployed a Checkbox Detection CNN model; boosted accuracy by 25% by adding layers, batch normalization, and dropout using TensorFlow (GPU with CUDA/cuDNN).
• Built and implemented a document classification system using TF-IDF and Support Vector Machines to accurately classify document types.
• Designed and integrated real-time hedge allocation models for the Mortgages business, enabling a dynamic response to market changes.

Jan 2019 – May 2019 · 4 mos · New York City Metropolitan Area

• - Tutoring for Prof Brian McFee. Leveraging the knowledge of Big Data Technologies.

Aug 2018 – Dec 2018 · 4 mos · New York City Metropolitan Area

• - Tutoring under Professor Adam Meyers for Python Programming Language. Leveraging the knowledge of Python, Programming Languages.

Jun 2018 – Aug 2018 · 2 mos · New York City Metropolitan Area

• Implemented Action Plans in Java to enhance the recommendation engine that suggests specific diet & supplements to address user’s health conditions.
• Optimized the knowledge base and database by writing evolutions scripts in SQL.
• Automated Integration Testing in shell script and wrote Unit Tests in Java to test the massive codebase.

Feb 2018 – May 2018 · 3 mos · New York City Metropolitan Area

• Researching on Big Data projects, analyzed data on Spark, Hive and Map Reduce on HPC cluster.
• Training and instructing students on Apache Spark, Apache Hive and Map Reduce.
• Currently instantiating OpenStack Cloud Computing environment.

Jul 2015 – Jun 2017 · 1 yr 11 mos · Mumbai Metropolitan Region

• Developed a capability “Twitter Sentiment Analysis” using Big Data Technologies – Apache Spark, Storm, Kafka and Camel by using Java and SQL database.
• As a Proof of Concept, individually modeled Machine Learning and NLP algorithms to evaluate factors causing prospect/opportunity losses in Java. Increased the efficiency by retraining the models by 44%.
• Individually automated the global taxonomy extraction as a Proof of Concept. Later, handled and drove the team as it became a full fledged product for the proprietary harmonization platform.
• Architected datasets and analyzed them on Big Data technologies such as Hive and Impala for further visualizations in Tableau. Automated the process on Oozie to provide recommendations for compensations and rebates.
• Designed components for proprietary product “Concordia” for data cleaning and mining.
• Optimized the framework from R to Hive to make it 11% faster for a Fortune 500 client.

Jan 2015 – Jun 2015 · 5 mos · Greater Bengaluru Area

• Modularized and rewrote the proprietary debugger (Jdebug) in Python used in Juniper OS; increasing the readability and efficiency.
• Programmed a tool “include-what-you-use” which detected and removed the libraries not being used, therefore making the execution faster by 8%.

Jun 2014 – Jul 2014 · 1 mo · Noida, Uttar Pradesh, India

• Implemented Event Management for the organization in Java.

Jun 2014 – Jul 2014 · 1 mo · Noida, Uttar Pradesh, India

• Implemented a student database management system in C.