Abstract
This section is dedicated to my core domain: Cloud Infrastructure, Platform Engineering, SRE, and Distributed Systems Design.
Here, I write about the architectural choices I make, why I choose specific tools or design patterns over others, the tradeoffs involved, and what happened in production:
- 🏛️ Architecture Decisions & Rationales: Deep dives into why a particular pattern (e.g. Hexagonal vs Layered, Async vs Multithreaded, Modular Monolith vs Microservices) was selected for a system.
- ☁️ Cloud & Platform Engineering: Kubernetes deployment architectures, infrastructure-as-code (Terraform), distributed state management, and high availability.
- ⚡ SRE & Resilience: Reliability engineering, observability architectures, failure domains, graceful degradations, and incident learnings.
- 🛠️ Software Engineering Craft: Async event-loop performance in Python, strict typing discipline, interface isolation, and non-blocking I/O.
📑 Featured Architecture Notes
- Zero-to-One MLOps (Part 1): Why Build by Hand and the 5-Phase Architecture — The infrastructure mismatch in ML, avoiding managed SaaS black boxes, and the 5-phase blueprint.
- Zero-to-One MLOps (Part 2): Hands-on Guide to k3s, MinIO Helm, and Traefik Ingress — Step-by-step bare metal k3s provisioning, decoupled Terraform loaders, and resolving 6 real infrastructure bugs.
- Why Async-First and Interface Abstraction for Real-Time Execution Engines — Solving blocking socket freezes, threadpool offloading via
asyncio.to_thread(), and backtest parity. - Modular Monolith vs Microservices for Financial Data Pipelines — Operational complexity vs modular domain isolation for multi-pillar market pipelines.