←
Mansi 👩‍💻5,658 posts
Mansi 👩‍💻 avatar

Mansi 👩‍💻

@MansiCodez

Senior Software Engineer | System Design | Interview guidance | AI. Collabs or DM : 📩 mansicodez@gmail.com for career guidance: topmate.io/mansicodez

◷ Joined July 2026▧ 417 Photos and videos
Mansi 👩‍💻 avatar
Mansi 👩‍💻@MansiCodez·Oct 18

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.

Mansi 👩‍💻 avatar
Mansi 👩‍💻@MansiCodez·14h

Recently I interviewed a Senior Engineer. Strong candidate. Kafka? ✅ Redis? ✅ Microservices? ✅ Distributed systems? ✅ Then I asked: "You're building the reservation system for a limited drop. Only 1 exclusive item left." Two buyers hit “Reserve Now” at the exact same millisecond. Both requests pass the “is available?” check. Both get a temporary reservation. Both try to confirm the order. Now you’ve sold the same unique item twice. How do you guarantee you never oversell — even with concurrent requests, client retries, network partitions, and partial failures? "Just use a database transaction." Doesn’t fully solve it if both requests read `available = 1` before either commits, or if the client never gets the response and retries. "Check availability then update." 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, the decrement succeeds but the order is never written, or the reservation expires while the payment is still processing? 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, GPU allocations, and limited resources are involved.

Mansi 👩‍💻 avatar
Mansi 👩‍💻@MansiCodez·15h

THE MOST DANGEROUS SOFTWARE ENGINEER IS NOT THE ONE WHO WRITES BAD CODE. It's the one who: -Deploys without testing. -Changes production without approval. -Ignores monitoring alerts. -Doesn't understand database migrations. -Pushes code without understanding it. -Blames other teams when things break. -Says "It works on my machine." Bad code can be refactored. Bad engineering decisions can take down an entire business.

Mansi 👩‍💻 avatar
Mansi 👩‍💻@MansiCodez·16h

WHY SOME SOFTWARE ENGINEERS GET PROMOTED FASTER: Engineer A: -Writes 2,000 lines of code. -Fixes 20 bugs. -Closes 30 Jira tickets. -Works late every day. Engineer B: -Identifies a major architectural bottleneck. -Reduces infrastructure costs by 30%. -Mentors 3 engineers. -Drives a cross-team project. -Prevents recurring production incidents. Engineer A is busy. Engineer B is creating leverage. Hard work matters. But impact gets you promoted.

Mansi 👩‍💻 avatar
Mansi 👩‍💻@MansiCodez·17h

20 TECHNOLOGIES EVERY BACKEND ENGINEER SHOULD KNOW: -Java -Python -Go -PostgreSQL -Redis -Kafka -RabbitMQ -Elasticsearch -Docker -Kubernetes -Nginx -gRPC -REST APIs -GraphQL -Prometheus -Grafana -Terraform -Linux -Git -AWS You don't need to master all 20. But understanding how these technologies fit together will make you a much stronger backend engineer. Save this roadmap.

Mansi 👩‍💻 avatar
Mansi 👩‍💻@MansiCodez·18h

TWO SOFTWARE ENGINEERS APPLY FOR THE SAME JOB. CANDIDATE A: -12 programming languages. -8 frameworks. -15 certifications. -10 years of experience. -5 impressive projects. CANDIDATE B: -Strong fundamentals. -Understands trade-offs. -Has solved complex production problems. -Can explain architectural decisions. -Knows how to debug unfamiliar systems. Who would you hire? A resume gets you the interview. Your ability to think gets you the job.

Mansi 👩‍💻 avatar
Mansi 👩‍💻@MansiCodez·19h

SENIOR ENGINEER: -What changed recently? -Is data being lost? -What's the blast radius? -Can we roll back safely? -How do we restore service? -How do we prevent this again? Experience isn't about knowing more commands. It's about making fewer dangerous decisions under pressure.

Mansi 👩‍💻 avatar
Mansi 👩‍💻@MansiCodez·19h

PRODUCTION DATABASE IS DOWN AT 2 AM. JUNIOR ENGINEER: -Restart the database. -Clear the cache. -Restart all services. -Check the logs. -Try random fixes. SENIOR ENGINEER:

Mansi 👩‍💻 avatar
Mansi 👩‍💻@MansiCodez·20h

EVERYONE WANTS TO BUILD AI AGENTS IN 2026. But very few understand what happens when an agent goes wrong. A production-ready AI agent needs: -Tool calling -Memory management -Context management -Retry mechanisms -Guardrails -Human approval -Observability -Cost management -Evaluation -Failure recovery Building an AI demo takes hours. Building a reliable AI product takes engineering.

Mansi 👩‍💻 avatar
Mansi 👩‍💻@MansiCodez·21h

INDIAN SOFTWARE ENGINEERS: Getting a ₹50 LPA offer feels like winning the lottery. But here's what nobody tells you: ₹30L fixed salary. ₹10L annual RSUs. ₹5L joining bonus. ₹5L performance bonus. Your actual monthly in-hand isn't ₹4.16 lakh. It can be significantly lower. CTC ≠ Fixed salary ≠ Take-home salary. Learn to negotiate your compensation structure, not just your CTC.

Mansi 👩‍💻 avatar
Mansi 👩‍💻@MansiCodez·22h

Great Engineer: -Understands the actual problem. -Writes simple, maintainable code. -Knows what NOT to build. -Debugs production issues calmly. -Makes the entire team better. Your value isn't measured by how much technology you know. It's measured by the problems you can solve.

Mansi 👩‍💻 avatar
Mansi 👩‍💻@MansiCodez·22h

THE DIFFERENCE BETWEEN A GOOD ENGINEER AND A GREAT ENGINEER: GOOD ENGINEER: -Knows 15 programming languages. -Has 50 certifications. -Knows every popular framework. -Can explain Kubernetes internals. -Has read 20 system design books. GREAT ENGINEER:

Mansi 👩‍💻 avatar
Mansi 👩‍💻@MansiCodez·23h

HOW TO BECOME AN AI ENGINEER IN 2026: A practical roadmap: Programming → Python, APIs, Git Fundamentals → ML basics, neural networks, embeddings LLMs → Transformers, tokens, context windows Prompt Engineering → System prompts, structured outputs RAG → Vector databases, chunking, retrieval AI Frameworks → LangChain, LlamaIndex, LangGraph AI Agents → Tool calling, workflows, multi-agent systems Evaluation → Hallucination testing, benchmarks, observability Deployment → Docker, APIs, cloud, monitoring Real Projects → Build AI applications people actually use. Don't spend 6 months just watching AI tutorials. Build something.

Mansi 👩‍💻 avatar
Mansi 👩‍💻@MansiCodez·23h

Bro asked Claude to count to 100,000. Claude turned it into a 3-hour engineering project, a threat, and a resignation letter 💀

I made Claude count to 100,000 and bro is already planning his brutal revenge 😭

Post attachmentPost attachment
Mansi 👩‍💻 avatar
Mansi 👩‍💻@MansiCodez·Nov 18

2026: Learn Python. Learn system design. Learn LLMs. Learn AI agents. Master AI coding tools. Build products using AI. Understand how AI systems fail. The programming language changed. The fundamentals didn't. Good engineers will always have an advantage.