Writing by Avinash Narwariyahttps://avnsh.me/https://avnsh.me/writing/self-healing-ai-systems/https://avnsh.me/writing/self-healing-ai-systems/How to build AI agent pipelines that recover on their own, using durable execution, verification loops, circuit breakers and observability.Sat, 22 Aug 2026 00:00:00 GMThttps://avnsh.me/writing/real-time-systems-at-scale/https://avnsh.me/writing/real-time-systems-at-scale/How real-time systems are built, where they break under load, and how to scale them with Kafka, WebSockets, backpressure and ordering guarantees.Wed, 17 Jun 2026 00:00:00 GMThttps://avnsh.me/writing/geography-of-speed-cdn/https://avnsh.me/writing/geography-of-speed-cdn/How CDNs work: edge caching, cutting latency, and three ways to add global asset delivery to a web app.Sun, 03 May 2026 00:00:00 GMThttps://avnsh.me/writing/redis-distributed-system-design/https://avnsh.me/writing/redis-distributed-system-design/Using Redis for more than caching: rate limiting, distributed locks, background jobs, and the trade-offs of each.Fri, 13 Feb 2026 00:00:00 GMThttps://avnsh.me/writing/real-time-voice-ai-architecture/https://avnsh.me/writing/real-time-voice-ai-architecture/How WebSockets, streaming speech-to-text, an LLM and text-to-speech fit together to keep voice AI conversations low latency.Tue, 03 Feb 2026 00:00:00 GMThttps://avnsh.me/writing/waf-practical-intro/https://avnsh.me/writing/waf-practical-intro/What a web application firewall is, where it sits in the stack, which common attacks it stops and how to deploy one.Tue, 20 Jan 2026 00:00:00 GMThttps://avnsh.me/writing/monolithic-first-auth-in-split-client-server-architecture/https://avnsh.me/writing/monolithic-first-auth-in-split-client-server-architecture/Why auth systems built for a monolith struggle once the client and server are split, and how a proxy bridges the gap.Fri, 16 Jan 2026 00:00:00 GMT