TaskFlow is an asynchronous task-processing platform I built to explore how background-job systems behave under load and how they can be deployed and operated reliably.

The API accepts tasks and returns HTTP 202 in under 10 ms, while Redis BLPOP workers process jobs independently in the background. During load testing, the system handled a spike of 1,000 tasks with zero dropped jobs.

I deployed the services on Kubernetes and configured HPA to scale workers from 2 to 50 replicas based on demand. Argo CD and Kustomize provide GitOps-based delivery, while GitHub Actions builds multi-stage Docker images. Prometheus and Grafana expose API, queue, worker, and cluster metrics.

Key engineering areas: asynchronous processing, queue workers, autoscaling, containerization, GitOps, CI/CD, and observability.

- [Application Repository](https://github.com/Tanveer-rajpurohit/taskflow)

- [Infrastructure Repository](https://github.com/Tanveer-rajpurohit/taskflow-infra)

- [Case Study](https://tanveer-singh.vercel.app/work/taskflow)