I'm Aamir a backend engineer who got tired of tutorials that stop at "Hello World." This is where I document what actually breaks in production and how I fix it.
I'm Aamir a backend engineer who got tired of tutorials that stop at "Hello World." This is where I document what actually breaks in production and how I fix it.
Open Source
Stop guessing which RAG pattern to use. Compare all 8 patterns with real benchmarks in one command — see which one wins for quality, latency, and cost on your actual documents.
$ rag-playbook compare --data ./docs/ --query "What is the refund policy?"
| Pattern | Latency | Cost | Chunks |
|---|---|---|---|
| naive * | 3,622ms | $0.0004 | 5 |
| parent_child | 5,056ms | $0.0005 | 5 |
| agentic | 5,008ms | $0.0005 | 5 |
| hybrid_search | 5,945ms | $0.0005 | 5 |
| self_correcting | 7,556ms | $0.0005 | 5 |
| hyde | 9,031ms | $0.0005 | 5 |
| reranking | 10,878ms | $0.0005 | 5 |
| query_decomp | 14,950ms | $0.0014 | 20 |
* Recommended: naive
Best quality/cost tradeoff. Latency: 3622ms, Cost: $0.0004
What I'm working on
A practical guide to preventing retry storms in distributed systems. Learn exponential backoff, circuit breakers, and jitter strategies that protect your services from cascading failures.
A practical guide to reducing LLM response validation latency. Learn parallel validation, tiered checking, caching strategies, and streaming validation to cut validation time from 400ms to 20ms.
A practical guide to sizing message queues and implementing backpressure. Learn how to prevent queue overflow, handle traffic spikes, and build systems that degrade gracefully under load.
Engineering Logs
FAQ
To share who I am and the kind of problems I've faced and solved over the years. Engineering can be isolating you solve something difficult, move on, and nobody else benefits from that struggle. If someone out there is stuck on the same problem I once was, maybe my experience can save them a few hours (or a few sleepless nights).
Have another question?