Execmind.ai - As the lead backend and DevOps engineer for ExecMind.ai’s MVP, I designed and deployed the core infrastructure powering real-time, multi-agent conversations. My work included integrating Stripe for payments and SendGrid for email, setting up role-based access control (RBAC) to keep data secure, and building REST APIs to support seamless interactions. To enable live chat, I used Django with Daphne and WebSockets, ensuring smooth, real-time communication.

On the DevOps side, I took full ownership, from domain setup and server provisioning to configuring Nginx and CI/CD pipelines, keeping everything running smoothly. Beyond the tech, I also led weekly client meetings to ensure our development aligned with business goals, bridging the gap between engineering and strategy.

Dronetjek.dk - At Dronetjek, I revamped the image upload flow in a monolithic Django application to handle high-resolution uploads more reliably. The original synchronous approach often caused delays and failures during upload, especially over slower networks.

To solve this, I integrated Celery with Redis to offload image processing tasks, such as resizing and format conversion using Pillow, to background workers. Once processed, images were uploaded to Amazon S3, and the corresponding database entries were updated with their URLs.

This change made the upload process significantly more reliable, cutting image processing delays by around 50% and reducing failure rates by over 70%. I also made minor improvements to the Nginx and Django configurations to better handle concurrent requests.

I worked as part of the Payments team at Amazon, contributing to Tier-1 and 2 microservices, critical to the payment ecosystem.

I contributed to the Prime Day 2024 peak readiness program, ensuring all microservices owned by my team were prepared for high traffic. During the event, I monitored one of Amazon’s most critical payment systems to ensure its stability.

I resolved a long-standing bug in a legacy Tier-1 service that caused order failures and financial accounting errors. This issue, which impacted financial accounting logic involving millions of dollars, required coordination with multiple teams to implement a fix ahead of Prime Day, preventing further disruptions and reducing manual recovery efforts.

I also made code changes across multiple services to enhance observability by improving service-level metrics. This led to a ~40% reduction in time-to-detect production issues and improved dashboard accuracy for on-call engineers by enabling finer-grained metrics.

I performed JDK migration for three production microservices, upgrading them from Java 8 to Java 17, updating dependencies and ensuring compatibility and system stability. Post-migration, the services saw up to a 15% improvement in throughput and a ~25% reduction in GC pause times under production load.

In addition, I led the deprecation of a Tier-2 microservice, coordinating with downstream teams and migrating functionality to a newer service. This effort resulted in a ~22% reduction in monthly operational costs.

Worked as a backend engineer on a platform connecting freelancers and companies with clients based on project history, feedback, and skill sets.

I designed and implemented systems for matching clients with suitable developers or companies for job postings, ensuring efficient and accurate pairing.

I worked with AWS org of Amazon, where I was working in Amazon’s Just Walk Store technology team. Our team was responsible for bringing every nuts and bolts together to launch a new JWO store.

There I developed an automation tool using Python, Boto3, and AWS CLI that reduced the weekly team effort by 60% through automation of AWS CFN stack creation/modifications, changeset execution, managing Route 53 entries, Lambda role additions, and dynamic markup table creation.

I also created a centralized dashboard to monitor 60+ microservices which each JWO store depended on, along with their dependencies, boosting team efficiency by 35% by preventing unnecessary configuration attempts.