Tag: #architecture
Writing on GPUs, LLMs, MLOps, Kubernetes — and mindset · 100 posts
100 posts are tagged #architecture; the most recent was published on 2026-08-15.
Often tagged together: #2026-03 53 #kubevirt 20 #microservices 11 #event-sourcing 10 #kubernetes 10
Most read in the last 90 days:
The Complete Guide to Database Caching Strategy: In the End, It's All Invalidation ♪ Listenable
Working from PostgreSQL 18, this guide builds caching up from the database outward: the cache layers that already exist — sharedbuffers and the OS cache — the computed cache that is a materialized view, the criteria for
2026-08-15 · 24 min read #database#postgresql#caching#performance-tuning#architectureWhy Langfuse Puts Traces in ClickHouse — How the Storage Layer Splits the Work
The first surprise when you self-host Langfuse is that there is not one datastore but four. ClickHouse, Postgres, Redis, and object storage each take a different share. Working from the official documentation, this post
2026-08-14 · 10 min read #observability#langfuse#clickhouse#architecture#postgresThe Plugin Kernel Architecture of DeepSeek Harness — What Makes an Agent You Can Rewind
Harness, the open source coding agent DeepSeek released, makes the model, the tools, and even the UI into plugins. But the interesting part is not the word plugin. It is two design decisions: representing every execution
2026-08-14 · 7 min read #ai-agent#architecture#plugin-system#observability#open-sourceWhat Replaces Technology — And How to Check Whether Your Stack Sits on That Trajectory
The final part of this series on open source projects that gave up their place. Cutting across the thirty-odd projects covered in the previous three parts, it distils five forces that replace technology: platform absorpt
2026-08-12 · 7 min read #open-source#architecture#license#migration#governanceYour Deploy Is the Load Test — What Happens When Nobody Designs the Cost of Seeding the Cache
A walk through how Canva moved the in-memory session revocation cache in its gateway from MySQL to S3. The problem was not the steady-state lookup cost but the startup cost of hundreds of pods seeding their caches simult
2026-08-09 · 8 min read #architecture#caching#s3#scalability#deploymentData Residency Is a Replication Topology, Not a Dropdown
Using the document Fastmail published when it opened an EU data region on August 3, 2026 as a textbook, this post lays out why the location of your data is not settled by picking a region once. The primary copy, the repl
2026-08-09 · 9 min read #architecture#data-residency#gdpr#replication#complianceThe Bauhaus: Fourteen Years That Changed a Hundred — How to Read the Shape of Things
The Bauhaus opened in Weimar in 1919 and closed in 1933 under Nazi pressure, so it existed for exactly fourteen years. It never held more than a few hundred students, the architecture department did not appear until eigh
2026-07-26 · 15 min read #humanities#design#architecture#design-history#modernismTurning a Customer Domain Into a Model — From Ubiquitous Language to Ontology
A customer's domain knowledge usually lives not in documents but as tacit knowledge inside people's heads. One core job of a Forward Deployed Engineer is turning that scattered tacit knowledge into an explicit model the
2026-07-15 · 11 min read #ontology#knowledge-graph#domain-driven-design#architecture#forward-deployed-engineerUpcycling Pretrained Checkpoints into Long-Context Hybrids — What HyLo Proposes
Hybrid sequence models (attention plus linear/SSM blocks) suit long context, but until now most had to be pretrained from scratch. The April 2026 preprint HyLo proposes a recipe to 'upcycle' an already-trained Transforme
2026-07-11 · 5 min read #ai#llm#long-context#architecture#efficiencyAnalyzing SOTA Multimodal LLMs — One Model to See, Hear, and Speak ♪ Listenable
How did a language model trained purely on text come to understand and generate images, audio, and video? This post walks through modality encoders and projectors, the unified token space, the any-to-any flow, native mul
2026-06-30 · 22 min read #multimodal-llm#any-to-any#vision-language#audio#architectureKubernetes Architecture Visualized — From the Control Plane to the Pod
We unpack the Kubernetes architecture with ASCII diagrams, from control-plane and node components to objects and controller loops, scheduling flow, networking, and storage. Request handling and the pod-creation sequence
2026-06-27 · 11 min read #kubernetes#architecture#control-plane#scheduling#networkingA Deep Dive into PostgreSQL Architecture — From MVCC to Vectors
A deep look at the internal architecture of PostgreSQL, from process and memory structure to MVCC, indexes, query execution, replication, partitioning, and extensions like pgvector. We unpack the core mechanisms with ASC
2026-06-27 · 11 min read #postgresql#database#mvcc#architecture#indexingVision LLM Architecture — How an Image Becomes Language ♪ Listenable
A vision-language model processes an image with a vision encoder, then passes it through a projector to produce tokens an LLM can read. From patch embedding to arbitrary-resolution handling, we trace the full path by whi
2026-06-26 · 20 min read #llm#vision-language-model#multimodal#vit#qwen2-vlArchitecture and Human Psychology — How Space Changes Us ♪ Listenable
We build buildings, but those buildings then shape us. Ceiling height changes our thinking, light and nature sway our mood, and the design of cities determines our happiness. Leaning on evidence, we explore how space aff
2026-06-21 · 27 min read #architecture#psychology#biophilia#urban-design#spaceOpen Banking and MyData API Architecture — The Technology of Financial Data Openness
The API architecture of Korean open banking shared network and MyData (personal credit information management). We cover standard API specs and auth tokens, comparison with UK Open Banking and the FAPI security profile,
2026-06-13 · 16 min read #finance#open-banking#mydata#api#oauth2Bank Deposit System Architecture — Dissecting the Heart of Core Banking ♪ Listenable
A deep dive into the deposit system at the heart of core banking. We cover account ledger design, deposit and withdrawal transaction processing, concurrency control, and how end-of-day closing coexists with 24x7 operatio
2026-06-13 · 19 min read #core-banking#deposit#ledger#architecture#bankingLending Systems and the LOS — Architecture from Loan Application to Disbursement ♪ Listenable
A deep dive into the overall architecture of lending systems at banks, card issuers, and capital companies, centered on the LOS (Loan Origination System). We walk through the pipeline from application to underwriting, ap
2026-06-13 · 19 min read #lending#los#loan-origination#core-banking#fintechBatch and End-of-Day Processing in Banking — The Architecture That Rules the Night
Why batch processing remains central in banking and the full picture of end-of-day (EOD) processing. We cover job schedulers and dependency DAGs, checkpoint restarts, partitioning and chunk processing, center-cut and dat
2026-06-13 · 18 min read #finance#batch#eod#spring-batch#schedulerBanking Ledger Design — Double-Entry Bookkeeping, Event Sourcing, and Consistency Down to the Last Cent
How to design the ledger at the heart of a banking system. We cover double-entry principles, the journal and posting data model, balance computation strategies, event sourcing, hot-account concurrency control, idempotenc
2026-06-13 · 17 min read #finance#ledger#double-entry#event-sourcing#core-bankingThe Rise of Diffusion LMs — Can They Become an Alternative to Autoregression
In June 2026 Google released DiffusionGemma, and text diffusion models are lighting up GeekNews and Hacker News. We critically analyze the structural limits of autoregressive generation, the principles of mask-based deno
2026-06-12 · 17 min read #llm#diffusion#text-generation#gemma#inference