All 25 articles, sorted alphabetically
Coding Assistants: How DeepSeek-V3 and Claude 3.5 Sonnet Became the New Standard for Software Engineering
The dream of "AI pair programming"—an intelligent assistant that truly understands code and helps engineers build better software faster—has long been a holy grail…
Read article →LoRA and QLoRA: Fine-tuning a Trillion-Parameter Model on a Single Home GPU
The era of massive foundation models—GPT-3, Llama 2/3, Mistral—has unleashed unprecedented AI capabilities.
Read article →Context Connect: A Word Context Game
Context Connect: A Word Context Game
Read article →A/B experimentation platform architecture
Deep-dive on building an A/B testing platform: deterministic salted assignment, exposure logging at the call site, the exposure-to-event join, a stats…
Read article →Real-time chat messaging architecture
Deep-dive on building a real-time chat system: persistent-connection gateways, a write path that persists and sequences before acknowledging, fan-out …
Read article →Dead-letter queues -- where failed messages go to be dealt with
Deep-dive on dead-letter queues (DLQs): the poison-message problem, sidelining after max retries, retry-then-dead-letter, preserving context (error/at…
Read article →Distributed counter architecture
Deep-dive on distributed counters: sharding to eliminate write contention, local aggregation and atomic adds, cached rollups for hot reads, HyperLogLo…
Read article →Distributed lock architecture
Deep-dive on distributed locking: a consensus-backed lock service granting a lease with a TTL that survives crashes, renewal heartbeats that extend it…
Read article →Geo proximity search architecture
Deep-dive on nearest-neighbour search: why B-trees cannot index 2D, space-filling curves from geohash to S2 Hilbert ordering, quadtree adaptivity, the…
Read article →Idempotency keys
Deep-dive on idempotency keys: client key generation and reuse, the idempotency store and first-vs-replay logic, concurrency locking, result caching, …
Read article →Inventory reservation architecture
Deep-dive on inventory reservation: the atomic check-and-decrement gate, available = on_hand − reserved, TTL-bounded holds and the reaper, confirm-on-…
Read article →News feed architecture
Deep-dive on building a social news feed at scale, where the central decision is when to do the work: fan-out on write (push a post id into every foll…
Read article →Object storage architecture
Deep-dive on object storage system design: the flat, immutable key-to-blob model that trades POSIX semantics for exabyte scale, how eleven-nines durab…
Read article →Transactional outbox -- reliable events with database writes
Deep-dive on the transactional outbox pattern: the dual-write problem, the outbox table and atomic same-transaction write, CDC vs polling relays, at-l…
Read article →Payment gateway architecture
Deep-dive on payment gateway architecture: idempotency keys, tokenization vaults and PCI scope, the payment state machine, risk and 3DS, immutable dou…
Read article →Priority queue service architecture
Deep-dive on priority queue services: why priority is a latency promise that only matters under contention, priority classes and ordering stores, the …
Read article →Saga Pattern Architecture in Depth
A 2500-word walkthrough of Saga pattern: choreography vs orchestration, steps + compensations, persistent state, idempotency, retries, observability.
Read article →Distributed job scheduler -- running work reliably at scale
Deep-dive on distributed job schedulers: why naive cron doesn't scale or survive failures, scheduling (cron/delays/intervals), distributed execut…
Read article →Typeahead architecture
Deep-dive on autocomplete systems: mining and scoring query logs, compiling FST indexes with per-node top-k, prefix-sharded in-memory serving, edge ca…
Read article →URL shortener architecture
Deep-dive on designing a URL shortener: the read/write skew and immutable mappings, counter+base62 vs KGS vs hashing key generation, cache-fronted red…
Read article →Webhook delivery platform architecture
Deep-dive on webhook delivery systems: durable ingest and subscription fan-out, per-endpoint circuit breakers, jittered exponential backoff, HMAC sign…
Read article →Workflow Orchestration Architecture in Depth
A 2500-word walkthrough of workflow orchestration: definition, scheduler, workers, durable state, task queue, retry, signals, child workflows, version…
Read article →Prompt Injection 101: How Hackers 'Jailbreak' AI and How to Defend Against It
Large Language Models (LLMs) are designed to be flexible, creative, and follow instructions presented in natural language.
Read article →Self-Healing Code: Using GenAI as a Smart Automation Layer for Legacy Codebase Migration
Nearly every mature enterprise is anchored by a "digital ball and chain": a critical, legacy codebase.
Read article →Tool Use (Function Calling): How Transformers Learned to Use Calculators, APIs, and Python
Large Language Models (LLMs) are masterful linguists. They can generate fluent, coherent text, summarize vast documents, and even craft creative prose.
Read article →