ROBERT4TECH

Tech Enthusiast | Software engineer

[ROBERT4TECH @Robert4tech]

@GalaxyFMUg Yam

[ROBERT4TECH @Robert4tech]

The difference is indentation (spacing before print).

[Tamale @256Rootyherman]

gonna build a toilet in the city where u enter for free but first watch ads to get out

[ROBERT4TECH @Robert4tech]

Afande, you are, right

[Muhoozi Kainerugaba @mkainerugaba]

Abaagala okunkuba oluseke like and retweet.

[ROBERT4TECH @Robert4tech]

@DaudiKabanda @mkainerugaba Yes twagala luseke😂😂😂

[ROBERT4TECH @Robert4tech]

Who were the Bachwezi

[ROBERT4TECH @Robert4tech]

If desktop is fast but mobile is slow, I’d throttle to Slow 4G + low CPU, then check JS bundle size, image optimization, render performance, and third-party scripts. Mobile bottlenecks are usually frontend + network constraints.

[Shravani @shrav_10]

Interviewer:
Your website loads fast on desktop but is extremely slow on mobile devices.
The backend looks fine.

Where would you start debugging?

[ROBERT4TECH @Robert4tech]

Feature flags. Same codebase, but behavior is conditionally enabled per user, % rollout, region, or tier. Enables gradual rollouts, A/B tests, and instant rollback no redeploy needed.

[Shreya @Oblivious9021]

How do platforms roll out features to selected users while using the SAME backend, SAME codebase, SAME APIs? 🤔

[ROBERT4TECH @Robert4tech]

😂😂😂😂

[Uganda Revenue Authority @URAuganda]

I promise 🙏
#FfeBano

[ROBERT4TECH @Robert4tech]

Database

[TheCyberPatronNetwork © @TheCyberPatron_]

What would be the target of an SQL injection attack?

a. database
b. DHCP
c. email
d. DNS

[ROBERT4TECH @Robert4tech]

🫡

[Daudi Kabanda @DaudiKabanda]

Who is prepared for the MK Fund??

[Daudi Kabanda @DaudiKabanda]

Who is prepared for the MK Fund??

[ROBERT4TECH @Robert4tech]

Am here sir

[ROBERT4TECH @Robert4tech]

Use replicas for reads with lag monitoring. Critical reads go to primary; non critical can tolerate slight staleness. Combine with caching, connection pools, and fallback to avoid overloading primary.

[THE CODE SCIENTIST @mysticwillz]

Your app uses a read-replica DB for performance.

One day, the replica lags behind by 15 seconds.
Users see outdated info.

You switch all reads to the primary and now it crashes from load.

How do you design a safe, resilient read strategy under failure?

[ROBERT4TECH @Robert4tech]

Use a write behind cache: Redis INCR for speed + durable queue/AOF for safety, then batch-flush counts to MySQL asynchronously. High throughput, no row lock storm, no data loss.

[Branko @brankopetric00]

You are designing a 'Like' button for a viral post.

Expected traffic: 100,000 likes per second.

Database: MySQL.

Problem: Updating a single row (`UPDATE posts SET likes = likes + 1 WHERE id = 1`) locks that row. All 100,000 requests serialize, killing performance.

Solution: You use Redis `INCR`.

But if Redis crashes, you lose the likes.

What 'Write-Behind' or 'Buffering' pattern allows you to capture high-speed writes in memory and persist them to SQL in batches safely?

[ROBERT4TECH @Robert4tech]

Design per client rate limits (API key/IP), use token bucket in Redis, enforce tier-based quotas, and dynamically tighten limits under high load. Return 429 with Retry After. Fairness > global throttling.

[THE CODE SCIENTIST @mysticwillz]

You’re running a public API.

Everything is fine…

Until one client sends 5,000 requests in 10 seconds.
Other users start timing out.

Your infra is under attack or just badly throttled.
How do you design dynamic rate limiting for fairness?

[ROBERT4TECH @Robert4tech]

If 10k users crash your app, it’s a scalability issue.
Fix with load balancing, horizontal scaling, Redis caching (sessions), DB optimization, queues for heavy tasks, and proper load testing before prod.

[kritika @vibeonX69]

Interviewer- Your product works perfectly in development but crashes when 10000 users log in at once.

Same code and Same features.

What would you use to fix this?

[kritika @vibeonX69]

Interviewer- Your product works perfectly in development but crashes when 10000 users log in at once.

Same code and Same features.

What would you use to fix this?

[ROBERT4TECH @Robert4tech]

You don’t stop users from breaking your system.
You design it assuming they will.
Validate inputs, enforce server side auth, rate limit, apply least privilege, and log everything. Defensive design > trusting users.

[Shreya @Oblivious9021]

Interviewer:
I asked 200+ engineers one simple question
How do you stop a user from breaking your system?

99% got it wrong!!