Microsoft SDE-2 Interview Experience | Selected

Harshit Kumar

Last Updated : 30 August 2026

213 Views

#interview

#microsoft

#dsa

#system-design

#sde-2

Adℹ

I recently gave the Microsoft SDE-2 interview and honestly it was a mix of DSA, some low-level/system design and a lot of focus on problem-solving. I'm just sharing my round-by-round experience here, hoping it helps anyone who's preparing for something similar.

Round 1 – Technical (DSA Problem)

Level: Medium
Duration: 60 mins
Mode: Remote

Question Asked : Shortest Distance of Open Cells to Nearest Mine

The interviewer gave me a maze represented as a grid of characters:

O → Open cell
X → Blocked cell
M → Mine

The task was to find the shortest distance of every open cell from its nearest mine.

Rules

Movement allowed only in 4 directions (up, down, left, right).
Distance of a mine to itself = 0.
Blocked and unreachable cells should return -1.

My Approach

The problem is essentially asking for the shortest distance from every open cell to the nearest mine.

A naive approach would be to start from each open cell and search for the closest mine, but that would be very inefficient.
Instead, the trick is to reverse the perspective:

Key Idea

Instead of starting from open cells, start the search from all the mines simultaneously.
This is where multi-source BFS comes into play.

Approach

1. Initialization

Scan the entire grid once.
For every mine (M), set its distance as 0 and push it into a queue.
For every blocked cell (X), mark it as -1 since it’s not traversable.
For all open cells (O), initialize them as -1 (unvisited).

2. BFS Propagation

Perform a Breadth-First Search (BFS) starting from all mines at once.
At each step:
- Pop a cell from the queue.
- Look at its 4 neighbors (up, down, left, right).
- If a neighbor is an open cell and hasn’t been visited yet:
  - Update its distance as current_distance + 1.
  - Push it into the queue.

3. Result

Because BFS explores in layers, the first time an open cell is reached, it is guaranteed to be from the nearest mine.
Every open cell automatically gets the shortest distance to its nearest mine.
Any cell that was never reached remains -1 (unreachable).

Why Multi-Source BFS Works Here

BFS naturally finds the shortest path in an unweighted grid.
By pushing all mines into the queue initially, we spread out from them simultaneously, so each open cell gets updated with the minimum distance to the closest mine in a single pass.
This avoids redundant searches and brings down the complexity to O(N × M) for an N × M grid.

Time Complexity

O(n × m), since each cell is visited at most once.

Code Snippet (what I wrote in the interview)

#include <bits/stdc++.h>
using namespace std;

void findShortestDistance(vector<vector<char>> &adj) {
    int n = adj.size();     // rows
    int m = adj[0].size();  // columns

    vector<vector<int>> dis(n, vector<int>(m, -1));
    queue<pair<int, int>> q;

    // Push all mines to the queue
    for (int i = 0; i < n; i++) {
        for (int j = 0; j < m; j++) {
            if (adj[i][j] == 'M') {
                q.push({i, j});
                dis[i][j] = 0;
            }
        }
    }

    // 4-directional movement
    int di[] = {0, 0, 1, -1};
    int dj[] = {1, -1, 0, 0};

    while (!q.empty()) {
        auto [row, col] = q.front();
        q.pop();

        for (int k = 0; k < 4; k++) {
            int newi = row + di[k];
            int newj = col + dj[k];

            if (newi >= 0 && newj >= 0 && newi < n && newj < m &&
                dis[newi][newj] == -1 && adj[newi][newj] == 'O') {
                dis[newi][newj] = dis[row][col] + 1;
                q.push({newi, newj});
            }
        }
    }

    // Print matrix as asked by interviewer
    for (int i = 0; i < n; i++) {
        for (int j = 0; j < m; j++) {
            cout << dis[i][j] << " ";
        }
        cout << endl;
    }
}

int main() {
    vector<vector<char>> adj = {
        {'O', 'M', 'O'},
        {'O', 'X', 'O'},
        {'O', 'O', 'O'}
    };

    findShortestDistance(adj);
    return 0;
}

Edge Cases I Discussed

Empty grid
Grid with no mines
Grid with all blocked cells
Multiple mines close to each other

Round 2 – DSA + Low-Level System Design

Level: Medium
Duration: 60 mins
Mode: Remote

Question Asked

The scenario was from a startup called BookNest, facing issues due to frequent DB hits for book previews. I had to design and implement a cache system that:

Stores a fixed number of book records.
Supports get(bookID) and put(bookID, bookData) operations.
Evicts the least recently used (LRU) book if the cache is full.
Both operations must run in O(1).

My Approach

The problem was essentially to design an LRU (Least Recently Used) Cache that can store book records and support get and put in constant time.
The key challenge was not just implementing caching, but also ensuring O(1) efficiency for both operations while handling evictions properly.

Approach

1. Identifying the Data Structures

I immediately thought of the classic LRU Cache design.

A HashMap would help in quickly locating any book record (bookID -> node) in constant time.
A Doubly Linked List would help in maintaining the order of usage, with the most recently used book at the head and the least recently used at the tail.

This combination ensures O(1) get/put while maintaining usage order.

2. How Operations Work

get(bookID)
- Look up in HashMap.
- If found, move the node to the head of the list (marking it as most recently used) and return the book.
- If not found, return -1.
put(bookID, bookData)
- If the book already exists, update its data and move it to the head.
- If it’s a new entry:
  - If the cache is full, evict the node at the tail (least recently used).
  - Insert the new book at the head and update the HashMap.

3. Complexity

Both operations (get and put) take O(1) because:

HashMap gives instant access.
Doubly Linked List allows quick node removals and insertions.

4. Abstraction & Clean Design

I wrapped this logic into a BookService interface and BookServiceImpl class to show abstraction, which the interviewer appreciated. This showed how caching could be integrated cleanly with a real-world service layer.

Why This Approach Works

Efficiency: Meets the strict O(1) requirement.
Scalability: Can handle large datasets as long as memory permits.
Real-world Relevance: Caching book previews reduces frequent DB hits, which was exactly the pain point in the BookNest scenario.

Code

#include <bits/stdc++.h>
using namespace std;

class Book {
public:
    int bookId;
    string author;

    Book(int id, const string &info) {
        bookId = id;
        author = info;
    }
};

class LRUCache {
private:
    int capacity;
    list<pair<int, Book>> accessList;
    unordered_map<int, list<pair<int, Book>>::iterator> cacheMap;
public:
    LRUCache(int cap) : capacity(cap) {}

    Book* get(int bookId) {
        if (cacheMap.find(bookId) == cacheMap.end()) return nullptr;

        accessList.splice(accessList.begin(), accessList, cacheMap[bookId]);
        return &(cacheMap[bookId]->second);
    }

    void put(int bookId, const string &bookData) {
        if (cacheMap.find(bookId) != cacheMap.end()) {
            cacheMap[bookId]->second = Book(bookId, bookData);
            accessList.splice(accessList.begin(), accessList, cacheMap[bookId]);
            return;
        }
        if (accessList.size() == capacity) {
            auto last = accessList.back();
            cacheMap.erase(last.first);
            accessList.pop_back();
        }
        accessList.push_front({bookId, Book(bookId, bookData)});
        cacheMap[bookId] = accessList.begin();
    }
};

Round 3 – System Design (Rate Limiter)

Level: Medium
Duration: 60 mins
Mode: Remote

Problem

In this round, I was asked to design a Rate Limiter, which is basically a system to control how many requests a user or client can make within a given time, so that APIs or services don't get abused. It is a classic system design problem.

Design a Rate Limiter that:

Limits number of API requests per user (e.g., 100 requests/min).
Works efficiently under high load.
Supports multiple users and scales well.

My Approach

I began by talking about the different rate limiting techniques such as

Fixed Window Counter
Sliding Window Log
Sliding Window Counter
Token Bucket Algorithm

Finally, I chose Token Bucket for its ability to handle burst traffic more smoothly.

Key Points I Discussed

Use a HashMap to store token counts per user.
Refill tokens at a fixed rate.
Reject requests if tokens are exhausted.
In distributed setups, Redis can be used to sync tokens across nodes.

Trade-offs & Scalability

Accuracy vs Performance: Sliding Window gives better accuracy, but Token Bucket provides better performance under high load.
Scalability: Extend across multiple servers using Redis or distributed coordination.

Round 4 – System Design (Bus Reservation Service)

Level: Medium
Duration: 60 mins
Mode: Remote

Problem

In this round, I had to design a bus reservation system for Microsoft employees. There were 25 fixed routes, each with a set of stops. In the morning, buses only picked up employees from their home stops and dropped them at the Microsoft campus, while in the evening the flow was reversed, boarding started from the campus and drop-offs happened at home stops.

Requirements

The problem requirements were to design a database schema that could handle passengers, drivers, bookings and routes. On top of that, I also had to design some basic APIs that a frontend application for employees could use.

My Approach

Database Design (Key Entities)

Passenger → passenger_id, name, start_stop, end_stop, date_of_travel
Driver → driver_id, name, assigned_bus_id
Bus → bus_id, route_id, capacity, assigned_driver_id
Route → route_id, list_of_stops
Stop → stop_id, location_name, geo_coordinates
Booking → booking_id, passenger_id, bus_id, status, trip_type

APIs

POST /register → Passenger signup
POST /login → Login
GET /routes/{route_id}/stops → View route stops
POST /booking → Create booking
GET /booking/active → View active bookings
GET /stops/nearby?lat=..&lng=.. → Nearby stops

Discussion Areas

Morning/evening rules validation.
Seat availability handling.
Extensibility for future features (e.g., notifications, live bus tracking, cancellations).

Round 5 : Hiring Manager Round

Level: Medium
Duration: 60 mins
Mode: Remote

This round was completely different from the coding and design ones. It wasn't about writing code or drawing system diagrams. it was more about me as a professional: how I think, how I handle pressure and how I approach situations that don't have straightforward answers.

What Happened

The manager literally walked through my resume line by line. Every project, every achievement, even small impact statements all were questioned.

For each point, I had to explain:

Why I did something the way I did
How exactly I achieved it
What happened after — was it successful, did it create an impact, did I learn something?

On top of that, I was given hypothetical scenarios:

“What if a key project is already failing and you’re brought in late — how would you handle it?”
“What if a senior architect strongly pushes back against your design — what would you do?”
“You join a brand-new domain/team with no prior experience — what’s your 30-day plan?”

These weren’t trick questions. The intent was clear: they wanted to see how I think under pressure, how I communicate and how I deal with ambiguity when there's no obvious right answer.

What They Were Really Testing

My ability to stay calm and think clearly under stressful situations.
How adaptable I am to new teams and unfamiliar domains.
Whether I can communicate ideas simply and effectively.
Most importantly, whether I take ownership — not just doing tasks, but driving impact and taking responsibility for outcomes.

Round 1 – Technical (DSA Problem)

Level: Medium
Duration: 60 mins
Mode: Remote

Question Asked : Shortest Distance of Open Cells to Nearest Mine

The interviewer gave me a maze represented as a grid of characters:

O → Open cell
X → Blocked cell
M → Mine

The task was to find the shortest distance of every open cell from its nearest mine.

Rules

Movement allowed only in 4 directions (up, down, left, right).
Distance of a mine to itself = 0.
Blocked and unreachable cells should return -1.

My Approach

The problem is essentially asking for the shortest distance from every open cell to the nearest mine.

A naive approach would be to start from each open cell and search for the closest mine, but that would be very inefficient.
\nInstead, the trick is to reverse the perspective:

Key Idea

Instead of starting from open cells, start the search from all the mines simultaneously.
\nThis is where multi-source BFS comes into play.

Approach

1. Initialization

Scan the entire grid once.
For every mine (M), set its distance as 0 and push it into a queue.
For every blocked cell (X), mark it as -1 since it’s not traversable.
For all open cells (O), initialize them as -1 (unvisited).

2. BFS Propagation

Perform a Breadth-First Search (BFS) starting from all mines at once.
At each step:\n
- Pop a cell from the queue.
- Look at its 4 neighbors (up, down, left, right).
- If a neighbor is an open cell and hasn’t been visited yet:\n
  - Update its distance as current_distance + 1.
  - Push it into the queue.
  \n
\n

3. Result

Because BFS explores in layers, the first time an open cell is reached, it is guaranteed to be from the nearest mine.
Every open cell automatically gets the shortest distance to its nearest mine.
Any cell that was never reached remains -1 (unreachable).

Why Multi-Source BFS Works Here

\n
BFS naturally finds the shortest path in an unweighted grid.
\n
\n
By pushing all mines into the queue initially, we spread out from them simultaneously,\nso each open cell gets updated with the minimum distance to the closest mine in a single pass.
\n
\n
This avoids redundant searches and brings down the complexity to O(N × M) for an N × M grid.
\n

Time Complexity

O(n × m), since each cell is visited at most once.

Code Snippet (what I wrote in the interview)

#include <bits/stdc++.h>\nusing namespace std;\n\nvoid findShortestDistance(vector<vector<char>> &adj) {\n    int n = adj.size();     // rows\n    int m = adj[0].size();  // columns\n\n    vector<vector<int>> dis(n, vector<int>(m, -1));\n    queue<pair<int, int>> q;\n\n    // Push all mines to the queue\n    for (int i = 0; i < n; i++) {\n        for (int j = 0; j < m; j++) {\n            if (adj[i][j] == 'M') {\n                q.push({i, j});\n                dis[i][j] = 0;\n            }\n        }\n    }\n\n    // 4-directional movement\n    int di[] = {0, 0, 1, -1};\n    int dj[] = {1, -1, 0, 0};\n\n    while (!q.empty()) {\n        auto [row, col] = q.front();\n        q.pop();\n\n        for (int k = 0; k < 4; k++) {\n            int newi = row + di[k];\n            int newj = col + dj[k];\n\n            if (newi >= 0 && newj >= 0 && newi < n && newj < m &&\n                dis[newi][newj] == -1 && adj[newi][newj] == 'O') {\n                dis[newi][newj] = dis[row][col] + 1;\n                q.push({newi, newj});\n            }\n        }\n    }\n\n    // Print matrix as asked by interviewer\n    for (int i = 0; i < n; i++) {\n        for (int j = 0; j < m; j++) {\n            cout << dis[i][j] << \" \";\n        }\n        cout << endl;\n    }\n}\n\nint main() {\n    vector<vector<char>> adj = {\n        {'O', 'M', 'O'},\n        {'O', 'X', 'O'},\n        {'O', 'O', 'O'}\n    };\n\n    findShortestDistance(adj);\n    return 0;\n}\n

Edge Cases I Discussed

Empty grid
Grid with no mines
Grid with all blocked cells
Multiple mines close to each other

Round 2 – DSA + Low-Level System Design

Level: Medium
\nDuration: 60 mins
\nMode: Remote

Question Asked

The scenario was from a startup called BookNest, facing issues due to frequent DB hits for book previews.\nI had to design and implement a cache system that:

Stores a fixed number of book records.
Supports get(bookID) and put(bookID, bookData) operations.
Evicts the least recently used (LRU) book if the cache is full.
Both operations must run in O(1).

My Approach

The problem was essentially to design an LRU (Least Recently Used) Cache that can store book records and support get and put in constant time.
\nThe key challenge was not just implementing caching, but also ensuring O(1) efficiency for both operations while handling evictions properly.

Approach

1. Identifying the Data Structures

I immediately thought of the classic LRU Cache design.

A HashMap would help in quickly locating any book record (bookID -> node) in constant time.
A Doubly Linked List would help in maintaining the order of usage, with the most recently used book at the head and the least recently used at the tail.

This combination ensures O(1) get/put while maintaining usage order.

2. How Operations Work

\n
get(bookID)
\n
- Look up in HashMap.
- If found, move the node to the head of the list (marking it as most recently used) and return the book.
- If not found, return -1.
\n
\n
put(bookID, bookData)
\n
- If the book already exists, update its data and move it to the head.
- If it’s a new entry:\n
  - If the cache is full, evict the node at the tail (least recently used).
  - Insert the new book at the head and update the HashMap.
  \n
\n

3. Complexity

Both operations (get and put) take O(1) because:

HashMap gives instant access.
Doubly Linked List allows quick node removals and insertions.

4. Abstraction & Clean Design

I wrapped this logic into a BookService interface and BookServiceImpl class to show abstraction, which the interviewer appreciated.\nThis showed how caching could be integrated cleanly with a real-world service layer.

Why This Approach Works

Efficiency: Meets the strict O(1) requirement.
Scalability: Can handle large datasets as long as memory permits.
Real-world Relevance: Caching book previews reduces frequent DB hits, which was exactly the pain point in the BookNest scenario.

Code

#include <bits/stdc++.h>\nusing namespace std;\n\nclass Book {\npublic:\n    int bookId;\n    string author;\n\n    Book(int id, const string &info) {\n        bookId = id;\n        author = info;\n    }\n};\n\nclass LRUCache {\nprivate:\n    int capacity;\n    list<pair<int, Book>> accessList;\n    unordered_map<int, list<pair<int, Book>>::iterator> cacheMap;\npublic:\n    LRUCache(int cap) : capacity(cap) {}\n\n    Book* get(int bookId) {\n        if (cacheMap.find(bookId) == cacheMap.end()) return nullptr;\n\n        accessList.splice(accessList.begin(), accessList, cacheMap[bookId]);\n        return &(cacheMap[bookId]->second);\n    }\n\n    void put(int bookId, const string &bookData) {\n        if (cacheMap.find(bookId) != cacheMap.end()) {\n            cacheMap[bookId]->second = Book(bookId, bookData);\n            accessList.splice(accessList.begin(), accessList, cacheMap[bookId]);\n            return;\n        }\n        if (accessList.size() == capacity) {\n            auto last = accessList.back();\n            cacheMap.erase(last.first);\n            accessList.pop_back();\n        }\n        accessList.push_front({bookId, Book(bookId, bookData)});\n        cacheMap[bookId] = accessList.begin();\n    }\n};\n\n

Round 3 – System Design (Rate Limiter)

Level: Medium
\nDuration: 60 mins
\nMode: Remote

Problem

In this round, I was asked to design a Rate Limiter, which is basically a system to control how many requests a user or client\ncan make within a given time, so that APIs or services don't get abused. It is a classic system design problem.

Design a Rate Limiter that:

Limits number of API requests per user (e.g., 100 requests/min).
Works efficiently under high load.
Supports multiple users and scales well.

My Approach

I began by talking about the different rate limiting techniques such as

Fixed Window Counter
Sliding Window Log
Sliding Window Counter
Token Bucket Algorithm

Finally, I chose Token Bucket for its ability to handle burst traffic more smoothly.

Key Points I Discussed

Use a HashMap to store token counts per user.
Refill tokens at a fixed rate.
Reject requests if tokens are exhausted.
In distributed setups, Redis can be used to sync tokens across nodes.

Trade-offs & Scalability

Accuracy vs Performance: Sliding Window gives better accuracy, but Token Bucket provides better performance under high load.
Scalability: Extend across multiple servers using Redis or distributed coordination.

Round 4 – System Design (Bus Reservation Service)

Level: Medium
\nDuration: 60 mins
\nMode: Remote

Problem

In this round, I had to design a bus reservation system for Microsoft employees. There were 25 fixed routes,\neach with a set of stops. In the morning, buses only picked up employees from their home stops and dropped them at the Microsoft campus,\nwhile in the evening the flow was reversed, boarding started from the campus and drop-offs happened at home stops.

Requirements

The problem requirements were to design a database schema that could handle passengers, drivers, bookings\nand routes. On top of that, I also had to design some basic APIs that a frontend application for employees could use.

My Approach

Database Design (Key Entities)

Passenger → passenger_id, name, start_stop, end_stop, date_of_travel
Driver → driver_id, name, assigned_bus_id
Bus → bus_id, route_id, capacity, assigned_driver_id
Route → route_id, list_of_stops
Stop → stop_id, location_name, geo_coordinates
Booking → booking_id, passenger_id, bus_id, status, trip_type

APIs

POST /register → Passenger signup
POST /login → Login
GET /routes/{route_id}/stops → View route stops
POST /booking → Create booking
GET /booking/active → View active bookings
GET /stops/nearby?lat=..&lng=.. → Nearby stops

Discussion Areas

Morning/evening rules validation.
Seat availability handling.
Extensibility for future features (e.g., notifications, live bus tracking, cancellations).

Round 5 : Hiring Manager Round

Level: Medium
\nDuration: 60 mins
\nMode: Remote

This round was completely different from the coding and design ones.\nIt wasn't about writing code or drawing system diagrams. it was more about me as a professional: how I think, how I handle pressure\nand how I approach situations that don't have straightforward answers.

What Happened

The manager literally walked through my resume line by line. Every project, every achievement, even small impact statements\nall were questioned.

For each point, I had to explain:

Why I did something the way I did
How exactly I achieved it
What happened after — was it successful, did it create an impact, did I learn something?

On top of that, I was given hypothetical scenarios:

“What if a key project is already failing and you’re brought in late — how would you handle it?”
“What if a senior architect strongly pushes back against your design — what would you do?”
“You join a brand-new domain/team with no prior experience — what’s your 30-day plan?”

These weren’t trick questions. The intent was clear: they wanted to see how I think under pressure, how I communicate\nand how I deal with ambiguity when there's no obvious right answer.

What They Were Really Testing

My ability to stay calm and think clearly under stressful situations.
How adaptable I am to new teams and unfamiliar domains.
Whether I can communicate ideas simply and effectively.
Most importantly, whether I take ownership — not just doing tasks, but driving impact and taking responsibility for outcomes.

","categoryId":363,"subCategoryId":142,"contentFormat":5,"blogId":17,"userId":"MjlfNF84LVs5XTJuLXRDX0FOZ2VfX0l0","title":"Microsoft SDE-2 Interview Experience | Selected","url":"microsoft-sde-2-interview-experience","bannerImage":"","seoDescription":"","generatedOn":"2025-09-14T06:41:02","updatedOn":"2026-08-30T06:49:33"},"popularContents":[{"id":172,"title":"Google L4 Software Engineer III interview experience | Selected","url":"google-l4-software-engineer-iii-interview-experience-selected","description":null,"seoDescription":"Google L4 Software Engineer III interview experience with API design, Googlyness, route matching, graph coding rounds, follow-up questions and tips.","contentType":5,"generatedOn":"2026-10-14T22:22:12.1484172Z","viewCount":0},{"id":166,"title":"System Design Interview – BIGGEST Mistakes to Avoid","url":"system-design-interview-biggest-mistakes-to-avoid","description":null,"seoDescription":"Learn the biggest mistakes to avoid in a system design interview, from unclear requirements to poor trade-off analysis. Includes real-world cases, advanced insights and examples to help you ace your next system design interview.","contentType":2,"generatedOn":"2026-10-14T22:22:12.1478611Z","viewCount":0},{"id":34,"title":"System Design Deep Dive: 25 Essential Interview Questions","url":"system-design-deep-dive-25-essential-interview-questions","description":null,"seoDescription":"","contentType":5,"generatedOn":"2026-10-14T22:22:12.1484032Z","viewCount":0},{"id":197,"title":"My Adobe SDE Interview Experience (Aug 2026) 5 Rounds, Selected","url":"my-adobe-sde-interview-experience-aug-2026-5-rounds-selected","description":null,"seoDescription":"Real Adobe SDE interview: 5 rounds, 4 DSA problems (Rotate Array, BSTs, Kth Smallest Matrix), code review & behavioral round. Full solutions + tips.","contentType":5,"generatedOn":"2026-10-14T22:22:12.1484205Z","viewCount":0},{"id":163,"title":"Top 50 SQL Interview Questions and Answers (Beginner to Intermediate)","url":"top-sql-interview-questions-and-answers-beginner-to-intermediate","description":null,"seoDescription":"Top 50 SQL Interview Questions and Answers (Beginner to Intermediate). Top SQL Interview Questions for fresher. Frequently asked SQL interview questions with answers","contentType":2,"generatedOn":"2026-10-14T22:22:12.1478529Z","viewCount":0},{"id":35,"title":"Cursor vs Copilot : Which AI Coding Assistant Wins in 2025?","url":"cursor-vs-copilot-which-ai-coding-assistant-truly-wins-in-2025","description":null,"seoDescription":"Compare Cursor vs GitHub Copilot on speed, context awareness, multi-file support, pricing and code quality to choose the best AI coding assistant for developers","contentType":5,"generatedOn":"2026-10-14T22:22:12.1484061Z","viewCount":0},{"id":167,"title":"Meesho SDE-1 Online Assessment Interview Experience (Remote) - Rejected","url":"meesho-sde-1-online-assessment-interview-experience-remote-rejected","description":null,"seoDescription":"Read a real Meesho SDE-1 online assessment experience with coding questions, DSA topics, solutions, difficulty level, preparation tips and interview insights.","contentType":5,"generatedOn":"2026-10-14T22:22:12.1484098Z","viewCount":0},{"id":173,"title":"Uber SDE-2 Interview Experience : LLD, HLD, DSA & Hiring Manager Round","url":"uber-sde-2-interview-experience-lld-hld-dsa-hiring-manager-round","description":null,"seoDescription":"Read a real Uber SDE-2 interview experience covering DSA, low-level design, high-level system design, behavioral rounds, hiring freeze and key takeaways.","contentType":5,"generatedOn":"2026-10-14T22:22:12.1484187Z","viewCount":0},{"id":135,"title":"Spinny SDE-1 Interview Experience (Selected)","url":"spinny-sde-1-interview-experience-selected","description":null,"seoDescription":"Spinny SDE-1 Interview Experience (Selected) | 4 Rounds Breakdown: Coding, LLD, Java, SQL + Real Interview Questions","contentType":5,"generatedOn":"2026-10-14T22:22:12.1484079Z","viewCount":0},{"id":171,"title":"Meesho SDE-1 Backend Interview Experience (Remote | 3 Round) - Rejected","url":"meesho-sde-1-backend-interview-experience-remote-3-round-rejected","description":null,"seoDescription":"Read a real Meesho SDE-1 Backend interview experience covering the online assessment, DSA interview, machine coding round, LLD questions and preparation tips.","contentType":5,"generatedOn":"2026-10-14T22:22:12.1484117Z","viewCount":0}],"latestContents":[{"id":219,"title":"6 Tools That Made My Life Easier as a Software Engineer","url":"6-tools-that-made-my-life-easier-as-a-software-engineer","description":null,"seoDescription":"The 6 developer tools I use daily as a software engineer: Git, VS Code, Docker, Postman, Chrome DevTools, and AI assistants, with real tips, mistakes, and trade","contentType":5,"generatedOn":"2026-10-14T22:22:12.148253Z","viewCount":0},{"id":218,"title":"When Should a Business Use Multiple AI Agents Instead of One?","url":"when-should-a-business-use-multiple-ai-agents-instead-of-one","description":null,"seoDescription":"Single agent or multi-agent AI? Learn the 4 signals that justify splitting agents by knowledge, permissions, and risk, plus routing and handoff tips.","contentType":5,"generatedOn":"2026-10-14T22:22:12.1482573Z","viewCount":0},{"id":216,"title":"AI in Clinical Trials Market Estimated to Experience a Hike in Growth by 2035","url":"ai-in-clinical-trials-market-estimated-to-experience-a-hike-in-growth-by-2035","description":null,"seoDescription":"The exclusive information about market dynamics serves as a valuable guide to predict economic scenarios and initiatives taken to enhance future growth. Our mar","contentType":5,"generatedOn":"2026-10-14T22:22:12.1482588Z","viewCount":0},{"id":217,"title":"AI in Clinical Trials Market Estimated to Experience a Hike in Growth by 2035","url":"ai-in-clinical-trials-market-estimated-to-experience-a-hike-in-growth-by-2035","description":null,"seoDescription":"The exclusive information about market dynamics serves as a valuable guide to predict economic scenarios and initiatives taken to enhance future growth. Our mar","contentType":5,"generatedOn":"2026-10-14T22:22:12.1482601Z","viewCount":0},{"id":215,"title":"What Happens When an AI Agent Doesn't Know the Answer?","url":"what-happens-when-an-ai-agent-doesnt-know-the-answer","description":null,"seoDescription":"Most AI agents fail quietly. They guess instead of saying \"I don't know.\" Here's how to design better fallback and human-handoff behavior.","contentType":5,"generatedOn":"2026-10-14T22:22:12.1482613Z","viewCount":0},{"id":214,"title":"Greedy Algorithms Explained: How They Work, When They Fail and How to Use Them","url":"greedy-algorithms-explained-how-they-work-when-they-fail-and-how-to-use-them","description":null,"seoDescription":"Learn greedy algorithms step by step with C# code Dijkstra, Huffman coding, knapsack and MSTs plus when greedy fails and how to prove it's correct","contentType":5,"generatedOn":"2026-10-14T22:22:12.1482629Z","viewCount":0},{"id":213,"title":"Amazon SDE II Interview Experience (Bangalore, Selected) ~ Aug 2026","url":"amazon-sde-ii-interview-experience-bangalore-selected-aug-2026","description":null,"seoDescription":"Amazon SDE II interview experience from Bangalore covering 4 rounds, DSA, LLD, system design, Leadership Principles, coding questions and selection.","contentType":5,"generatedOn":"2026-10-14T22:22:12.1482673Z","viewCount":0},{"id":212,"title":"How to Get 10x Better AI Answers Without Writing Better Prompts : 10 Proven Techniques","url":"how-to-get-10x-better-ai-answers-without-writing-better-prompts-10-proven-techniques","description":null,"seoDescription":"Get better AI answers without complex prompts. Learn 10 practical techniques using context, examples, tools, feedback, and verification.","contentType":5,"generatedOn":"2026-10-14T22:22:12.1482688Z","viewCount":0},{"id":211,"title":"I Interviewed for a Microsoft SDE I Role. Here's Everything That Happened.","url":"i-interviewed-for-a-microsoft-sde-i-role-heres-everything-that-happened","description":null,"seoDescription":"My real Microsoft SDE I interview experience, all 4 rounds, the exact DSA problems with fully tested solutions, the WhatsApp system design round","contentType":5,"generatedOn":"2026-10-14T22:22:12.1482701Z","viewCount":0},{"id":140,"title":"PolicyBazaar Software Development Engineer (SDE) Interview Experience | DSA, Core CS & System Design","url":"policybazaar-software-development-engineer-sde-interview-experience-dsa-core-cs-system-design","description":null,"seoDescription":"Detailed PolicyBazaar SDE interview experience covering 3 rounds: DSA coding, core CS assessment and system design with key learnings and insights.","contentType":5,"generatedOn":"2026-10-14T22:22:12.1482721Z","viewCount":0}],"relatedContents":[{"id":211,"title":"I Interviewed for a Microsoft SDE I Role. Here's Everything That Happened.","url":"i-interviewed-for-a-microsoft-sde-i-role-heres-everything-that-happened","description":null,"seoDescription":"My real Microsoft SDE I interview experience, all 4 rounds, the exact DSA problems with fully tested solutions, the WhatsApp system design round","contentType":5,"generatedOn":"2026-10-14T22:22:12.1518157Z","viewCount":0},{"id":183,"title":"My Microsoft SWE Interview Experience (July 2026): 4 Rounds, Real Mistakes and the Offer","url":"my-microsoft-swe-interview-experience-july-2026-4-rounds-real-mistakes-and-the-offer","description":null,"seoDescription":"Read my Microsoft Software Engineer (L62) interview experience with coding, LLD, HLD, managerial rounds, mistakes, solutions, and preparation tips.","contentType":5,"generatedOn":"2026-10-14T22:22:12.1518194Z","viewCount":0},{"id":190,"title":"My Microsoft SDE-2 Interview Experience (2026) : Paused, Closed and Then Reopened Two Months Later","url":"my-microsoft-sde-2-interview-experience-2026-paused-closed-and-then-reopened-two-months-later","description":null,"seoDescription":"Microsoft SDE-2 interview experience covering 5 rounds, DSA, AI-assisted LLD, HLD, OS, C++, coding questions, hiring process and key learnings.","contentType":5,"generatedOn":"2026-10-14T22:22:12.1518211Z","viewCount":0},{"id":213,"title":"Amazon SDE II Interview Experience (Bangalore, Selected) ~ Aug 2026","url":"amazon-sde-ii-interview-experience-bangalore-selected-aug-2026","description":null,"seoDescription":"Amazon SDE II interview experience from Bangalore covering 4 rounds, DSA, LLD, system design, Leadership Principles, coding questions and selection.","contentType":5,"generatedOn":"2026-10-14T22:22:12.1518257Z","viewCount":0},{"id":139,"title":"Blinkit (Eternal) SDE-1 Interview Experience (June 2026) – Offer Received","url":"blinkit-eternal-sde-1-interview-experience-june-2026-offer-received","description":null,"seoDescription":"Blinkit SDE-1 interview experience with a 30 LPA offer. Covers House Robber DP, ticket booking system design, culture fit round and tips.","contentType":5,"generatedOn":"2026-10-14T22:22:12.1518273Z","viewCount":0},{"id":137,"title":"JPMorgan SDE-2 Interview Experience - 3 Rounds | Java, Code Review & System Design","url":"jpmorgan-sde-2-interview-experience-3-rounds-java-code-review-system-design","description":null,"seoDescription":" JPMorgan Chase SDE-2 interview experience covering code review, Core Java, system design, concurrency, behavioral rounds and key learnings.","contentType":5,"generatedOn":"2026-10-14T22:22:12.1518289Z","viewCount":0},{"id":140,"title":"PolicyBazaar Software Development Engineer (SDE) Interview Experience | DSA, Core CS & System Design","url":"policybazaar-software-development-engineer-sde-interview-experience-dsa-core-cs-system-design","description":null,"seoDescription":"Detailed PolicyBazaar SDE interview experience covering 3 rounds: DSA coding, core CS assessment and system design with key learnings and insights.","contentType":5,"generatedOn":"2026-10-14T22:22:12.1518302Z","viewCount":0},{"id":135,"title":"Spinny SDE-1 Interview Experience (Selected)","url":"spinny-sde-1-interview-experience-selected","description":null,"seoDescription":"Spinny SDE-1 Interview Experience (Selected) | 4 Rounds Breakdown: Coding, LLD, Java, SQL + Real Interview Questions","contentType":5,"generatedOn":"2026-10-14T22:22:12.1518315Z","viewCount":0},{"id":209,"title":"Adobe Computer Scientist 1 Interview Experience: Real Questions, Solutions & Tips","url":"adobe-computer-scientist-1-interview-experience-real-questions-solutions-tips","description":null,"seoDescription":"Real Adobe CS-1 interview experience covering C++, DSA questions, coding solutions, reschedules, rejection, interview challenges and preparation tips.","contentType":5,"generatedOn":"2026-10-14T22:22:12.1518331Z","viewCount":0}]}}},"source":{"isMobile":false}}; window.__CLIENT_RENDER__ = false;

Microsoft SDE-2 Interview Experience | Selected

Round 1 – Technical (DSA Problem)

Question Asked : Shortest Distance of Open Cells to Nearest Mine

Rules

My Approach

Key Idea

Approach

1. Initialization

2. BFS Propagation

3. Result

Why Multi-Source BFS Works Here

Time Complexity

Code Snippet (what I wrote in the interview)

Edge Cases I Discussed

Round 2 – DSA + Low-Level System Design

Question Asked

My Approach

Approach

1. Identifying the Data Structures

2. How Operations Work

3. Complexity

4. Abstraction & Clean Design

Why This Approach Works

Code

Round 3 – System Design (Rate Limiter)

Problem

My Approach

Key Points I Discussed

Trade-offs & Scalability

Round 4 – System Design (Bus Reservation Service)

Problem

Requirements

My Approach

Database Design (Key Entities)

APIs

Discussion Areas

Round 5 : Hiring Manager Round

What Happened

What They Were Really Testing

Trending Developer Reads

Round 1 – Technical (DSA Problem)

Question Asked : Shortest Distance of Open Cells to Nearest Mine

Rules

My Approach

Key Idea

Approach

1. Initialization

2. BFS Propagation

3. Result

Why Multi-Source BFS Works Here

Time Complexity

Code Snippet (what I wrote in the interview)

Edge Cases I Discussed

Round 2 – DSA + Low-Level System Design

Question Asked

My Approach

Approach

1. Identifying the Data Structures

2. How Operations Work

3. Complexity

4. Abstraction & Clean Design

Why This Approach Works

Code

Round 3 – System Design (Rate Limiter)

Problem

My Approach

Key Points I Discussed

Trade-offs & Scalability

Round 4 – System Design (Bus Reservation Service)

Problem

Requirements

My Approach

Database Design (Key Entities)

APIs

Discussion Areas

Round 5 : Hiring Manager Round

What Happened

What They Were Really Testing