Senior Software Engineer | System Design | Interview guidance | AI. Collabs or DM : 📩 mansicodez@gmail.com for career guidance: https://xitter.twitterapi.xyz/t.co/Mcz9YSfysI
Joined July 2026
- Tweets5.2K
- Following1.9K
- Followers8.4K
- Likes11.5K
Pinned Tweet
Recently I interviewed a Senior Engineer.
Strong candidate.
Kafka? ✅
Redis? ✅
Microservices? ✅
Distributed systems? ✅
Then I asked:
"You're building the inventory service for a flash sale.
Only 1 item left in stock."
Two users hit “Buy Now” at the exact same millisecond.
Both requests pass the “is available?” check.
Both create an order.
Both deduct stock.
Now you’ve sold 2 items when only 1 existed.
How do you guarantee you never oversell — even with concurrent requests, client retries, and partial failures?
"Just use a database transaction."
Doesn’t fully solve it if both requests read `stock = 1` before either writes, or if the response is lost and the client retries.
"Check stock before deducting."
What happens when both checks happen at the same time?
"Use Redis DECR and reject if it goes negative."
What happens if Redis is down, or the decrement succeeds but the order is never persisted?
Your turn:
How would you design this inventory reservation API so concurrent buyers and retries never oversell the last item?
There’s a specific distributed-systems concept behind this.
And you’ll encounter it everywhere inventory, tickets, seats, coupons, and limited resources are involved.
Recently I interviewed a Senior Engineer.
Strong candidate.
Kafka? ✅
Redis? ✅
Microservices? ✅
Distributed systems? ✅
Then I asked:
“You’re building the system that generates unique short links (like bit.ly).
Users are creating thousands of links per second.”
Two requests arrive at the exact same millisecond.
Both generate the same short code “x7k9p”.
Both try to insert it into the database.
Both succeed in writing the mapping.
Now two completely different long URLs point to the same short link.
How do you guarantee every short code is unique — even with concurrent requests, retries, and multiple service instances?
“Just generate a UUID.”
UUIDs are long and ugly. Users want short, memorable codes.
“Check if the code exists, then insert if not.”
What happens when both checks see “does not exist” at the same time?
“Use Redis SETNX.”
What happens if Redis is temporarily unavailable, or the set succeeds but the DB write fails?
Your turn:
How would you design this short-link generation service so concurrent requests never produce duplicate codes?
There’s a specific distributed-systems concept behind this.
And you’ll hit it every time you need guaranteed uniqueness under concurrency (usernames, referral codes, coupon codes, order IDs, etc.).
2
4
374
Unpopular opinion:
AI won’t replace software engineers.
It will replace engineers who only know how to write code and can’t explain system design, trade-offs, or why something should exist.
7
12
580
If the interviewer says “design Twitter”
and you start drawing boxes immediately, never do this:
- Jump straight into microservices
- Start with the database
- Say “I’d use Kafka”
Instead say:
3
7
1,095
Unpopular opinion:
We’re burning tokens on coding and dev work. The real use of AI is curing cancer and other diseases.
2
219
Mid-level engineer: “We should use Kafka for this.”
Senior engineer: “What’s the actual throughput and what happens if this fails?”
The biggest jump in interviews isn’t more knowledge.
It’s asking better questions.
1
7
1,037
If the interviewer asks “How would you design Instagram?”
and you immediately start talking about microservices, you’ve already lost.
Most candidates fail because they design before they understand the problem.
Spend the first few minutes asking about scale and what actually matters.
6
405
THE DIFFERENCE BETWEEN A MID-LEVEL AND SENIOR ENGINEER IN INTERVIEWS:
Mid-level: “I’ll use Redis for caching.”
Senior: “What are the read/write patterns and what’s the acceptable staleness?”
Stop naming tools. Start asking better questions.
4
15
2,306
Unpopular opinion:
Most engineers fail system design interviews not because they don’t know the concepts.
They fail because they start designing before they understand the requirements.
Always spend the first 5 minutes clarifying scale, constraints, and what actually matters.
6
8
497
BREAKING: Nvidia CEO Jensen Huang faces removal calls after second affair report.
WSJ says board preparing to oust him following NYT report of a relationship with VP Kimberly Powell (who reports to him), on top of his admitted affair with Alice Huang — an executive U.S. officials suspect is a Chinese spy. Stock has tanked, investigations are underway, and succession risk is now front and center.
@grok confirm this news
1
6
719
