ทวีตที่ปักหมุด
Madara
33.8K posts
I had to try this! I love it 😁😁
How is this guy's?
Amira Zairi@azed_ai
Would you buy this chair? Prompt 👇
English
Madara รีทวีตแล้ว
📂 DSA Roadmap to Fix the Top 10 Beginner Mistakes
┃
┣ 🚫📂 1. Solving Without Understanding
┃ ┣ 📂 Read problem properly
┃ ┣ 📂 Dry run examples
┃ ┗ 📂 Understand WHY solution works
┃
┣ 🚫📂 2. Jumping to Hard Problems
┃ ┣ 📂 Start with easy
┃ ┣ 📂 Build patterns
┃ ┗ 📂 Gradually increase difficulty
┃
┣ 🎯📂 Solve THESE first (Foundation Set)
┃ ┣ 📂 1 — Two Sum
┃ ┣ 📂 20 — Valid Parentheses
┃ ┣ 📂 53 — Maximum Subarray
┃ ┣ 📂 121 — Best Time to Buy & Sell Stock
┃ ┗ 📂 125 — Valid Palindrome
┃
┣ 🚫📂 3. No Revision
┃ ┣ 📂 Revisit solved questions
┃ ┣ 📂 Maintain notes
┃ ┗ 📂 Weekly revision
┃
┣ 🚫📂 4. Copy-Pasting Solutions
┃ ┣ 📂 Try for 20–30 mins
┃ ┣ 📂 Understand solution
┃ ┗ 📂 Re-code without help
┃
┣ 🚫📂 5. Ignoring Complexity
┃ ┣ 📂 Learn Big-O
┃ ┣ 📂 Optimize brute force
┃ ┗ 📂 Compare approaches
┃
┣ 🚫📂 6. No Consistency
┃ ┣ 📂 Daily 2–3 problems
┃ ┣ 📂 Fixed schedule
┃ ┗ 📂 Track progress
┃
┣ 🚫📂 7. Not Learning Patterns
┃ ┣ 📂 Sliding window
┃ ┣ 📂 Two pointers
┃ ┣ 📂 Binary search
┃ ┗ 📂 DP basics
┃
┣ 🚫📂 8. Avoiding Weak Topics
┃ ┣ 📂 Identify weak areas
┃ ┣ 📂 Practice extra
┃ ┗ 📂 Face hard topics
┃
┣ 🚫📂 9. Too Much Tutorial Watching
┃ ┣ 📂 Limit passive learning
┃ ┣ 📂 Solve more than watch
┃ ┗ 📂 Learn by doing
┃
┣ 🚫📂 10. No Mock Practice
┃ ┣ 📂 Timed practice
┃ ┣ 📂 Simulate interviews
┃ ┗ 📂 Analyze mistakes
Follow this for 2–3 months consistently.
Are you solving questions…
or actually improving?
English
🚨Been grinding this for a whole week straight and finally cracked it 🔥
You follow people → they follow back Then they quietly unfollow you but keep you following them😪
Or you follow and get zero follow-backs
I was doing this shit manually on Android every week… hours gone.
So I built a dead-simple Inspect Element script. Just open X on Chrome → paste → instantly see who doesn’t follow you back + recent unfollowers.
No apps. No login. No risk.
Script + full instructions below👇
github.com/adityadand/twi…
If this saves you way lot of time
⭐if you like repo will bring new scripts too
English
- Claude = coding. ($20/mo)
- Supabase = backend. (Free)
- Vercel = deploying. (Free)
- Namecheap = domain. ($12/yr)
- Stripe = payments. (2.9%/transaction)
- GitHub = version control. (Free)
- Resend = emails. (Free)
- Clerk = auth. (Free)
- Cloudflare = DNS. (Free)
- PostHog = analytics. (Free)
- Sentry = error tracking. (Free)
- Upstash = Redis. (Free)
- Pinecone = vector DB. (Free)
Total monthly cost to run a startup: ~$20
There has never been a cheaper time to build.
English
As an AI Infrastructure Engineer.
Please learn:
- GPU/VRAM fundamentals, quantization & batching
- vLLM / TensorRT-LLM / inference optimization
- KV caching, speculative decoding & token throughput
- Distributed training basics (DDP/FSDP/DeepSpeed)
- Model serving & autoscaling
- Vector DB retrieval pipelines
- Prompt caching & cost optimization
- Observability for LLM apps
This is what production AI teams actually care about.
English
🧵 Day 11/30 — #LLD
DIP (Dependency Inversion Principle)
👉 High-level modules should NOT depend on low-level modules
👉 Both should depend on abstractions
In simple terms:
→ Don’t depend on concrete classes
→ Depend on interfaces / contracts
---
Why this matters 👇
When DIP is violated:
→ Tight coupling between components
→ Hard to change implementations
→ Difficult to test (no mocking)
→ Small changes ripple across system
---
Classic example 👇
❌ Bad design
OrderService directly depends on MySQLDatabase
→ Changing DB = changing service
→ Testing becomes hard
---
✅ Better design
→ Create Database interface
→ MySQL, Postgres implement it
→ OrderService depends on interface
Now:
→ Swap DB without changing service
→ Easy to test (mock interface)
---
💡 Key idea:
Business logic should NOT care about:
→ Database
→ External APIs
→ Frameworks
It should depend on abstractions only
---
⚠️ Tradeoffs:
→ More interfaces
→ More setup initially
→ Slightly complex structure
But…
→ Highly flexible system
→ Easy to extend & test
→ Loosely coupled architecture
---
🔥 Rule:
If changing a low-level detail breaks your core logic →
You’re violating DIP
---
Where this shows up 👇
→ Database switching
→ Payment gateway integration
→ Notification providers
→ External APIs
---
Would you:
A) Directly depend on concrete classes
B) Use abstractions for flexibility
What would you choose and why?
👇
Ajit Gupta@unfiltered_ajit
🧵 Day 10/30 — #LLD ISP (Interface Segregation Principle) 👉 “Clients should not be forced to depend on methods they don’t use” In simple terms: → Don’t create fat interfaces → Create small, focused ones --- Why this matters 👇 When ISP is violated: → Classes implement unused methods → Leads to empty / dummy implementations → Increases confusion → Makes code harder to maintain --- Classic example 👇 ❌ Bad design One interface: Worker → work() → eat() Now a Robot implements it 🤖 → work() ✅ → eat()? ❌ (doesn’t make sense) Forced implementation = bad design --- ✅ Better design Split interfaces: → Workable → work() → Eatable → eat() Now: → Human → implements both → Robot → implements only work() Clean. Logical. Flexible. --- 💡 Key idea: Interfaces should be: → Small → Specific → Role-based Not “one-size-fits-all” --- ⚠️ Tradeoffs: → More interfaces → Slightly more design effort But… → Better flexibility → Easier testing → Cleaner architecture --- 🔥 Rule: If a class is implementing methods it doesn’t need → Your interface is too big --- Where this shows up 👇 → Payment systems (different capabilities) → Notification services (SMS vs Email) → API design (modular endpoints) --- Would you: A) One big interface for everything B) Multiple small focused interfaces What would you choose and why? 👇
English
