Interview experience
HSBC Interview Questions and Answers (2026)
Real HSBC interview questions - candidate interview experiences and HR round prep, in one place.
Overview
Section titled “Overview”HSBC’s fresher technology hiring runs a two-phase online test, a genuinely eliminating HSBC Values Assessment, a fundamentals-focused technical round, and an HR round - distinct from its separate banking-operations and analytics hiring tracks.
HSBC interview process at a glance
Section titled “HSBC interview process at a glance”| Round | Typical duration | What they test |
|---|---|---|
| Online Assessment - Phase 1 | ~75 min | ~20 aptitude, ~20 English, ~30 technical MCQs (DBMS, OOP, OS/Linux) |
| Online Assessment - Phase 2 | ~30 min | 2 coding problems, easy to easy-moderate; pass at least one fully |
| HSBC Values Assessment | 20-30 min | Situational-judgement behavioural test; an elimination stage |
| Technical Interview | 20-30 min | Data structures, implementations, SQL/NoSQL, ACID, projects |
| HR Interview | ~30 min | Background, motivation, HSBC values, location preference, awareness |
One widely shared on-campus account tracked the funnel as 254 candidates at the assessment, 76 at the values assessment, 47 at the technical interview and 15 at the HR round - so the online assessment does the heavy filtering.
Common HR questions asked at HSBC
Section titled “Common HR questions asked at HSBC”- Tell me about yourself and your family background.
- Why HSBC, and what do you know about our values?
- Which location would you prefer, and are you comfortable relocating?
- If you had to automate one thing in your college, what would it be?
- What is the difference between confidence and overconfidence?
Sample answer frameworks for each of these are on the HSBC HR interview questions page.
Common technical interview questions and answers
Section titled “Common technical interview questions and answers”Q: How would you implement a stack using an array versus a linked list?
With an array you keep an index called top starting at -1; push increments top and writes the value, pop reads and decrements. It gives O(1) push and pop, excellent cache locality, but a fixed capacity unless you grow the array by doubling, which costs O(n) occasionally and O(1) amortised. With a singly linked list you push by creating a node and pointing it at the current head, and pop by moving the head to head.next - also O(1), with no capacity limit and no resize copies, but with a pointer of memory overhead per element and poor cache behaviour. HSBC interviewers commonly ask this exact comparison, so lead with the complexity being identical and the difference being memory layout and resizing.
Q: What are the ACID properties of a transaction?
Atomicity means a transaction is all-or-nothing - if the debit succeeds but the credit fails, the whole thing rolls back. Consistency means the database moves from one valid state to another, respecting every constraint, key, and trigger. Isolation means concurrent transactions do not see each other’s intermediate state; the isolation level chosen (Read Uncommitted, Read Committed, Repeatable Read, Serializable) decides which anomalies - dirty reads, non-repeatable reads, phantom reads - remain possible. Durability means once COMMIT returns, the change survives a crash, which engines achieve with a write-ahead log flushed to disk before the commit is acknowledged.
Q: When would you choose SQL over NoSQL?
Choose a relational database when the data is highly structured and relational, when you need multi-row multi-table transactions with strong ACID guarantees, and when ad-hoc querying and joins matter - which is why core banking ledgers are almost always relational. Choose NoSQL when the schema is fluid or varies per document, when you need horizontal scale-out across many nodes beyond what a single relational primary can serve, or when the access pattern is a simple key lookup at very high throughput. The genuine trade-off is that most distributed NoSQL stores relax consistency to gain availability and partition tolerance, so a read can return slightly stale data - unacceptable for an account balance, fine for a product catalogue or activity feed.
Q: What is a heap, and what are the complexities of its operations?
A heap is a complete binary tree stored in an array, satisfying the heap property: in a min-heap every parent is less than or equal to its children, so the minimum sits at index 0. For a node at index i the children are at 2i+1 and 2i+2 and the parent is at (i-1)/2, so no pointers are needed. Insert appends at the end and sifts up in O(log n); extract-min swaps the root with the last element, shrinks the array, and sifts down, also O(log n); peeking at the minimum is O(1). Building a heap from an existing array of n elements is O(n), not O(n log n), which is a favourite follow-up.
Q: What is the difference between a process and a thread?
A process is an independent execution unit with its own virtual address space, file descriptor table, and resources; a thread is a unit of execution inside a process that shares that address space and those descriptors with its sibling threads but has its own stack, registers, and program counter. Because threads share memory, communication between them is cheap - just shared variables - but that is also why they need mutexes and other synchronisation to avoid race conditions, while processes are isolated and must use pipes, sockets, or shared memory to talk. Context switching between threads is cheaper than between processes because the page tables and TLB do not have to be swapped.
Q: How does a hash map work, and how does it handle collisions?
A hash map applies a hash function to the key to produce an index into an array of buckets, giving average O(1) insert, lookup, and delete. Collisions - two keys landing in the same bucket - are resolved either by chaining, where each bucket holds a linked list (or, above a threshold in modern Java, a balanced tree), or by open addressing, where the map probes to the next free slot linearly or quadratically. Worst-case lookup degrades to O(n) with chaining if every key collides, or O(log n) once a bucket is treeified. When the load factor - entries divided by buckets - crosses a threshold such as 0.75, the map resizes and rehashes every entry, which is why an unexpected O(n) pause can show up on a single put.
Q: What is the difference between an array and a linked list?
An array stores elements contiguously, so indexing is O(1) by address arithmetic and iteration is cache-friendly, but inserting or deleting in the middle is O(n) because the remaining elements must shift, and the size is fixed unless you reallocate. A linked list stores each element in a node holding a pointer to the next, so inserting or deleting is O(1) once you already hold the reference to the position, but reaching position k costs O(k) since you must traverse. Memory-wise, the array wastes any unused capacity while the linked list pays a pointer per element plus allocator overhead. The practical rule: pick an array for read-heavy indexed access, a linked list only when you do frequent insert or delete at a known position.
Q: Which Linux commands would you use to investigate a slow or stuck process?
Start with top or htop to see which process is consuming CPU and memory, and ps -ef or ps aux piped through grep to confirm the PID and full command line. Use df -h and free -m to rule out a full disk or exhausted memory, since those cause symptoms that look like application slowness. For a specific PID, lsof -p shows the open files and sockets it holds, and strace -p shows the system calls it is currently making, which quickly reveals whether it is blocked on I/O or spinning. Tail the application log with tail -f, and use grep -n on the log for the error signature before concluding anything.
Frequently asked questions about HSBC interviews
Section titled “Frequently asked questions about HSBC interviews”What is the HSBC interview process for freshers?
HSBC Technology India’s Trainee Software Engineer / SDE hiring is commonly reported as four stages: 1. Online Assessment in two phases - an aptitude, English and technical MCQ paper (a widely shared pattern is 70 questions in 75 minutes: 20 aptitude, 20 English, 30 technical MCQs on DBMS, OOP and OS/Linux), followed by a coding phase of 2 problems in about 30 minutes. 2. HSBC Values Assessment - a situational-judgement style behavioural test that is itself an elimination stage. 3. Technical Interview - roughly 20-30 minutes on data structures, DBMS and projects. 4. HR Interview - about 30 minutes on background, motivation, HSBC’s values and location preference.
What is the HSBC Values Assessment?
It is a behavioural or situational-judgement assessment that sits between the coding test and the interviews, and candidates report it as a genuine elimination round rather than a formality. You are shown workplace scenarios and asked to choose or rank responses. HSBC publishes its values - including dependability, openness, connectedness and doing the right thing - so read them on the careers site before you sit the assessment and answer consistently with them rather than second-guessing.
What coding questions does HSBC ask?
The online coding phase is reported as 2 problems in around 30 minutes, typically one easy and one easy-to-moderate, with candidates needing to pass at least one fully to progress. Topics stay in the standard fresher range - arrays, strings, hashing and basic logic - rather than hard algorithmic puzzles. In the technical interview, candidates are more often asked to explain data structures and implementations (for example, implementing a stack using an array versus a linked list) than to solve a fresh judged problem.
Is the HSBC technical interview hard?
Candidates describe it as short and fundamentals-focused rather than difficult - typically 20-30 minutes covering the data structures you know (heaps, stacks, queues, linked lists, arrays), how you would implement them, SQL versus NoSQL and ACID properties, your preferred programming language, and a walkthrough of your projects and academics. Depth on fundamentals matters more than competitive-programming range.