Aswin Shine's learning library

I started as a UI designer and got more curious about what happens after the deploy button than before it. Now I do both: I design the screens, build the app, and run it on AWS so it stays up. That means one person from the Figma frame to the Grafana dashboard. Some of what Ive shipped: • Penwave, a blogging platform I designed and ran on EC2 with Docker Compose, then rebuilt on EKS with ArgoCD and Helm. A version tag starts a pipeline with ten gates, and a failed health check rolls back. • An F1 race-engineer chatbot on RAG (LangChain, Pinecone, GPT-4.1-mini). Its Docker image started at 21 GB. I got it to about 5. • A Chrome extension that reads a YouTube videos comments and tells you how the audience feels: LightGBM served from MLflows registry, API on EC2, 85.3% accuracy. • Nullify, an incident tool whose README claimed 10,000 signals a second. I measured about 525, then batched the writes to about 17,400. Where I can help: • Getting an app from a laptop to AWS with Terraform, Docker and CI/CD • Kubernetes on EKS, GitOps with ArgoCD, monitoring with Prometheus and Grafana • Taking an ML or LLM feature from a notebook to an API people can call • Designing the interface in Figma before any of it I'd rather have boring releases and numbers I measured myself. Every project on aswin-shine.dev has a what broke section with the commits behind the fixes. Based in Jaipur, working with teams anywhere. Need something designed, built and shipped? Message me here. Specialties: AWS, EKS, Terraform, Kubernetes, ArgoCD, Helm, GitHub Actions, Docker, Prometheus, Grafana, MLflow, LangChain, RAG, React, Next.js, Figma