Hey, I'm Divij Gera
I
A Full Stack Developer building and managing the Front-end and Back-end of Websites and Web Applications that leads to the success of the overall product. Check out my work below.
About Me
Here you will find more information about me, what I do, and my current skills mostly in terms of programming and technology
Get to know me!
I'm a Software Engineer currently working at Microsoft on the Azure Resource Notifications team, building pub-sub services handling hundreds of billions of notifications daily.
Previously at Amazon, I contributed to Lumos, a cloud-based payment solution handling $2B+ transactions. I have extensive experience with AWS cloud technologies, microservices architecture, and building scalable distributed systems.
I'm also passionate about Deep Learning and Computer Vision, with a published paper at IEEE ICME 2023 on adversarial robustness in neural networks.
My Skills
Experience
Here you will find details about my professional journey and the companies I've had the privilege to work with
Microsoft
Software Engineer 2
February 2025 - Present
Microsoft
Software Engineer 2
February 2025 - Present
Working as part of the Azure Resource Notifications team in Azure Platform and Management Services, building a pub-sub service handling massive scale notifications.
Key Achievements
- ●Building pub-sub service handling 100B+ notifications/day on publisher end and 500B+/day on subscriber end
- ●Developed microservice using C# .NET for dynamic service management and capacity/throttling estimations
- ●Reduced weekly on-call time by 4+ hours through automation and improved tooling
Technologies
Microsoft
February 2025 - PresentAmazon
Software Development Engineer
September 2022 - February 2025
Amazon
Software Development Engineer
September 2022 - February 2025
Contributed to Lumos, a cloud-based solution for Amazon Payments handling $2B+ payments across Amazon's ecosystem.
Key Achievements
- ●Engineered infrastructure provisioning using AWS CDK and CloudFormation across 10+ AWS services
- ●Designed event-driven system using Publisher-Subscriber model with AWS Kinesis, DynamoDB, and SES
- ●Achieved 40% annual cost reduction ($120K) by optimizing ECS service instances
- ●Built React UI for lambda service with API Gateway, saving 2+ hours of weekly developer time
- ●Resolved 50+ high-severity incidents while on-call, reducing team operational load
Technologies
Amazon
September 2022 - February 2025Projects
Here you will find some of the personal and clients projects that I created with each project containing its own case study
Meta Perturbed Re-Id Defense
A novel approach to adversarial robustness in person re-identification systems. Introduced anisotropic and isotropic perturbations to design a stochastic neural network, and derived a novel feature covariance alignment (FCA) loss achieving high clean performance while providing robustness against different attacks. Published at IEEE ICME 2023.
Read Publication