Everything your app needs beyond CRUD.
CRUD is the easy part. Production means caching, queues, databases at scale, failures, security, and now AI. I'm Ashwani Kumar, a backend engineer with ten years of shipping, and I break down the system design behind all of it, plus what's new in tech and AI. Free, for every developer.
CRUD is 10%. Here's the other 90%.
- Caching
- What to cache, for how long, and what happens on a stampede.
- Queues & events
- Retries, idempotency, dead letters, and ordering.
- Databases
- Indexes, migrations, replicas, and the query that loads the table.
- Scaling
- Load balancing, sharding, rate limits, and noisy neighbours.
- Reliability
- Timeouts, backoff, circuit breakers, and graceful shutdown.
- Observability
- Logs, metrics, traces, and alerts that wake the right person.
- Security
- Auth, secrets, supply chain, and what an agent can reach.
- AI in production
- LLM calls as dependencies: cost, latency, fallbacks, evals.
- 27→88%
- of KYC checks fully automated
- RentoMojo
- 2h→15s
- bulk matching job, rewritten
- PeopleGrove
- +150%
- daily active users after shipping Mel, an AI assistant
- Radius Agent
- 1000s
- of long-lived headless Chrome sessions kept running
- browserless
Latest writing
all posts →How Dead Letter Exchanges Prevent Distributed Systems From Silently Failing
One bad payment event can keep thousands of healthy events waiting. The interesting part is not what a DLX is - it is deciding which failures should leave the hot path, and what happens after they do.
- 21 Oct 2026note
One word keeps your table writable while an index builds
CREATE INDEX vs CREATE INDEX CONCURRENTLY in Postgres, and the four things that bite people when they switch.
#postgres#migrations2 min→ - 20 Oct 2026note
An AI agent just got a 503 from your API
A human gives up after two refreshes. An agent doesn't. Four things I'd ship on any API that agents will call.
#api-design#ai-agents#reliability2 min→
newsletter
Get the next failure mode by email.
One essay at a time, with the code, plus the checklists I make. Straight to your inbox.
What you'll find here
System design
Essays and short notes on how production systems are built and why they break.
explore →What's new
New tech and new AI, with what it actually changes for the systems you run.
explore →Teardowns
New apps taken apart: how they work, how you'd build your own, and how they found users.
explore →There's an AI for that
AI tools worth a developer's time, and the system design under them.
explore →There's open source for that
Paid apps and their free twins, and how to self-host them.
explore →your turn
Got a system you want explained, or an app you want taken apart?
Send it over. The best questions become the next post, free for everyone.