The Practical Stack

Most engineers use LLMs daily but can't explain what happens between prompt and response. #LLM #LargeLanguageModels #Transformers #GenerativeAI #SoftwareEngineering #MLEngineering #AIEngineering #BackendEngineering #ThePracticalStack

Most engineers use LLMs daily but can't explain what happens between prompt and response. Here's the full pipeline — 8 concepts, one image. Save this. 🔁 Tokenization → Embeddings → Attention → Transformer → KV Cache → Sampling → Context Window → Inference Optimisation

Spring Security Login- Error Handling and Localization is.gd/oXDZ38 #java #springmvc #springframework #rest #springboot

#Java NIO2 File Attributes javadevjournal.com/java/java-nio2… #springframework #rest #springboot #springmvc

Spring Method Security is.gd/zSdzQM #rest #springmvc #springframework #java #springboot

#Java Access Modifiers is.gd/8z7ktC #rest #springmvc #springboot #springframework

Pick your cache strategy based on what you’re willing to lose: Freshness Latency Durability Everything else is implementation detail.

Most engineers pick a caching strategy based on what they've used before — not what the system actually needs. 8 strategies. Different tradeoffs. Cache-aside ≠ always the answer. 🧵Which one do you default to ? #DistributedSystems #SystemDesign

ArrayList contains() Method is.gd/mLUvrP #springframework #springmvc #rest #java #springboot

Prototype Design Pattern is.gd/qZaYr0 #rest #springboot #springframework #springmvc #java

Meta replaced softmax with custom activations in GDPA kernels. Result: 2x forward pass speedup, 30% training throughput gain. thepracticalstack461.substack.com/publish/post/1… #MachineLearning #PyTorch

How to Integrate Spring Boot with Hibernate is.gd/8BJwzm #java #springboot #springmvc #rest #springframework

AVL Tree is.gd/pR0fZM #springmvc #springframework #java #rest #springboot

Grab lost 40% of eng time to repetitive queries. They didn't build a bigger model. They decomposed it like microservices. Schema drift hits one agent. Not the whole system. Boundaries matter more than models. The Practical Stack 👇 thepracticalstack461.substack.com #DistributedSystems

Tier-1 bank: sub-5ms p99 at 1.6M msg/sec with strict ordering intact. The unlock wasn't just tuning. Multi-DC replica placement done right — skip this and tail latency spikes under load. The Practical Stack 👇 thepracticalstack461.substack.com

Your K8s scheduler is blind to GPU utilization. A job at 3% looks identical to one that finished. Utilization-aware preemption fixes this — but the idle threshold is the entire bet. The Practical Stack 👇 thepracticalstack461.substack.com #Kubernetes #GPUComputing

Spring Security Authentication is.gd/6MznDg #springboot #springframework #rest #java #springmvc

#REST Web Services Bascis is.gd/286l1r #java #springframework #springboot #springmvc

@ppetryszen Cleaner primitives help, but the real question is how this behaves under pressure. Scheduling looks simple—until resource contention forces tradeoffs.

K8s v1.35 shipped numeric toleration operators. SLA-based scheduling native to the control plane. No webhooks. No external controllers. thepracticalstack461.substack.com/publish/post/1… #Kubernetes #CloudNative #SRE