• Led the backend development of a new product utilizing microservices architecture to generate smart notes for healthcare professionals based on patient-doctor conversations. Technologies: Python, Docker, Jenkins, GCP services, and Gemini LLM.

• Engineered a distributed locking mechanism using Redis to prevent race condition in a note due to concurrent microservice executions.

• Improved on reliability, observability and alerting of all the microservices by enhancing code, logging and cloud policies.

• Developed a Vector Search microservice to enhance keyword matching capabilities. Tech: Python, Flask and ML Encoder model.

• Benchmarked, identified and resolved performance bottlenecks within an microservice, resulting in a 25% reduction in runtime.

• Contributed to DevOps migration of old infrastructure from AWS to GCP, ensuring zero downtime for all services.

• Designed Distributed System, developed REST APIs & rolled out Probability-based throttling feature which solved 90% of over throttling issues with the customers.

• Setup end-to-end VPCE service for new regions in Kinesis using Java, Python, AWS CloudFormation, LB & EC2 to provide more secure way for users to call Kinesis service.

• Leaded onboarding of Kinesis fleet with 30,000+ hosts to Managed Fleet service which reduced 20% of SEV2 tickets per week related to host maintenance.

• Migrated service from long term credentials (Odin MS) to short term rotating credentials (Turtle) to improve security of the Kinesis services.

• Ensuring System reliability by mitigating issues (50+ weekly SEV2; 100+ pages) during 24/7 on-call shifts, to ensure customer retention.

• Mentored new joiners in team to reduce strain on senior engineers.

• Worked on developing backed for an NLP-based application using Java, Spring, MongoDB, AWS CloudFront & Cognito service to generate annotation results for provided documents using Machine Learning model.

• Developed an end-to-end cloud solution using Python, AWS S3, EC2 & RDS to process real-time sensor data.

• Developed Python & Spark-based data pipeline in Azure Databricks platform to filter out regression data.

• Contributed to Expertiza (open-source project) by extending the functionality in Ruby on Rails – helped to implement the functionality of making a review public.

• Analyzed MVC architecture of the code and identified suboptimized & dead code.

• Refactored codebase to reduce cognitive, cyclomatic, ABC & Lines of Code complexities.

• Collaborated with Engineering team for understanding Aircraft Engine Data and discovering insights from it.

• Created an ETL pipeline using Python, SQL & Microsoft Azure services and reduced the processing time by 50%.

• Performed feature engineering using R & Python and narrowed down data to multivariate time series data.

• Developed Deep Learning Models (LSTM) and Statistical Models (ARIMA) for processed aircraft engine data for predictive component failure and predicted accurately with RMSE value of less than 0.1.

• Visualized the processed and predicted engine data using Microsoft Excel and Visual Basic.