Accenture Custom Software Engineer Interview Questions 2026: 30 Questions with Answers

"Custom Software Engineer" confuses a lot of applicants because it is not a technology; it is Accenture's name for the job family that builds bespoke applications for clients rather than configuring packaged products. The interview is therefore a general software-engineering interview anchored on the primary skill in the job description, with Accenture's own emphasis on client delivery, agile and cloud. This page explains the role and the 2026 process, then gives 30 questions with sample answers across coding, OOP, SQL, web, cloud, delivery scenarios and HR.

Most openings under this title in 2026 ask for zero to two years of experience at Associate or Analyst level, with separate postings for Senior Analyst and above. If you are a fresher, the first four groups are your syllabus; if you are a lateral, the delivery and cloud groups decide the offer.

How the Accenture Custom Software Engineer process works in 2026

As of September 2026, candidate reports and 2026 hiring guides describe three steps for off-campus and lateral Custom Software Engineering openings.

Step 1: online assessment. A coding or aptitude test on HackerEarth or Accenture's own assessment portal. Typical content is logical reasoning, basic programming MCQs and one or two coding problems in the language of your choice; some drives add a communication assessment. Elimination is strict at this step.

Step 2: technical interview. Forty-five to sixty minutes with a senior engineer or team lead. Your primary skill (Java, Python, .NET or a front-end framework) in depth, data structures, SQL, web and API concepts, your project, and one or two live coding tasks. For laterals, questions about your current project's architecture and delivery practices.

Step 3: HR interview. Background, motivation, location and shift flexibility, notice period and compensation. Accenture's HR round is short but the answers on flexibility carry weight.

Campus freshers. On-campus and pooled-campus candidates go through Accenture's fresher assessments and interviews first and are then mapped to job families including Custom Software Engineering; the technical questions below still apply.

Timeline and difficulty. Glassdoor's aggregate for the Custom Software Engineering Analyst title shows an average of 35 days to hire, a difficulty of 3 out of 5, and 67 percent positive experiences. Locations in 2026 postings include Hyderabad, Bengaluru, Pune, Chennai and Mumbai.

Compensation. Not published on the job posts and dependent on level, city and skill. Confirm the level in your offer and compare it for that level; do not negotiate against a number from a comment thread.

What the Accenture interviewer is testing

• Primary skill depth. If the JD says Java, expect collections, OOP, exceptions, multithreading basics and Spring; if Python, data structures, OOP, decorators, generators, and a framework like Django or FastAPI.
• Problem solving. One or two medium problems, spoken through before coding.
• Data and APIs. SQL by hand, REST design, JSON handling.
• Cloud and DevOps vocabulary. Git branching, CI/CD, containers, one cloud provider's basics.
• Client-delivery mindset. Agile ceremonies, estimation, handling change, communicating status.
• Flexibility. Location, shifts, and working on whatever stack the project needs.

30 Accenture Custom Software Engineer interview questions with sample answers

Coding and problem solving

1. Find the first non-repeating character in a string.

"Two passes: build a frequency map of characters, then scan the string again and return the first character with count one. O(n) time, O(k) space for the alphabet. A single pass with an ordered map also works. I would confirm whether case matters and what to return when every character repeats."

2. Rotate an array to the right by k positions in place.

"Reverse the whole array, then reverse the first k elements, then reverse the rest. k modulo n handles k larger than n. O(n) time, O(1) extra space. The naive approach of shifting one step k times is O(nk) and the interviewer will ask for better."

3. Given a list of transactions, find the customer with the highest total spend.

"Group by customer ID summing the amount, then take the maximum. In Python a dictionary accumulates totals in one pass; in Java a HashMap<String, Double> with merge. O(n). If the data does not fit in memory, I would say this is a group-by in SQL or a Spark job. Accenture panels like the 'what if it is large' follow-up."

4. Implement a queue using two stacks.

"Push onto stack one for enqueue. For dequeue, if stack two is empty, pop everything from stack one into stack two, then pop from stack two. Each element moves at most twice, so operations are amortised O(1)."

5. Check whether brackets in an expression are balanced.

"A stack: push opening brackets, on a closing bracket pop and check it matches; at the end the stack must be empty. O(n). Edge cases: a closing bracket with an empty stack, and unmatched openers left at the end."

6. What is the time complexity of common operations on an array, linked list, hash map and balanced BST?

"Array: O(1) index, O(n) insert or delete in the middle. Linked list: O(1) insert or delete with a node reference, O(n) search. Hash map: O(1) average for get and put, O(n) worst case with bad hashing. Balanced BST: O(log n) for search, insert and delete, and it keeps order, which a hash map does not."

Primary language and OOP

7. Explain the four pillars of OOP with examples from your project.

Tie each to your code. "Encapsulation: the Account class exposed withdraw() and kept balance private. Abstraction: NotificationService hid whether email or SMS was used. Inheritance: SavingsAccount extended Account. Polymorphism: calculateInterest() behaved differently per account type." Definitions alone do not pass an Accenture technical round.

8. Java: what is the difference between HashMap, LinkedHashMap and TreeMap?

"HashMap has no order and O(1) average operations. LinkedHashMap keeps insertion or access order and is what an LRU cache uses. TreeMap keeps keys sorted with O(log n) operations and supports range queries such as headMap. I choose by whether I need order and which kind."

9. Python: what are decorators and generators? Where did you use them?

"A decorator wraps a function to add behaviour; I used one to log execution time on API handlers. A generator yields values lazily, so I processed a large CSV line by line without loading it into memory. Both are questions Accenture asks for Python-primary Custom Software Engineer openings."

10. What is exception handling best practice in your language?

"Catch specific exceptions, not the base class. Never swallow silently; log with context. Use finally or try-with-resources and with blocks for cleanup. Wrap low-level exceptions into domain exceptions at layer boundaries so callers get meaningful errors. Let unrecoverable errors propagate to a global handler."

11. Explain interfaces versus abstract classes and give a design use.

"An interface defines a contract with no state; an abstract class can carry state and shared code. I define PaymentGateway as an interface so tests can mock it and Razorpay or a bank API can be swapped, and I use an abstract BaseReport when three report types share formatting logic."

12. What is multithreading, and what problems does it introduce?

"Running several threads in one process for concurrency, useful for I/O-bound work like calling multiple services in parallel. Problems: race conditions on shared mutable state, deadlocks when locks are acquired in different orders, and visibility issues. I use thread pools and immutable data where possible, and synchronise only the smallest critical sections."

SQL and data

13. Write a query to find the second-highest salary.

"SELECT MAX(salary) FROM employees WHERE salary < (SELECT MAX(salary) FROM employees); or with DENSE_RANK() over salary descending and picking rank two, which extends naturally to the nth highest and per-department variants."

14. Inner join versus left join, with an example where the choice changes the result.

"Inner join returns rows with matches on both sides; left join keeps every row from the left table with nulls where there is no match. Listing customers and their orders: inner join drops customers with no orders; left join keeps them with null order columns, which is what a 'customers who never ordered' report needs, filtered with WHERE o.id IS NULL."

15. What is an index, and when does adding one make things worse?

"A data structure, usually a B-tree, that lets the database find rows without scanning the table. It slows inserts, updates and deletes because the index must change too, wastes space on low-selectivity columns, and is ignored if the query wraps the column in a function. I index foreign keys and frequent filter columns, and check the execution plan."

16. Explain normalisation and when you would denormalise.

"Normalisation removes redundancy by splitting data into related tables: 1NF atomic values, 2NF full dependency on the key, 3NF no transitive dependencies. I denormalise deliberately for reporting or read-heavy screens, storing a computed total or a name copy, and keep it consistent through the write path."

17. How do you prevent SQL injection?

"Parameterised queries or an ORM, never string concatenation with user input. Least-privilege database accounts, input validation as a second layer, and no error messages that expose the query. I would also mention that stored procedures alone do not prevent it if they build dynamic SQL inside."

Web, APIs and cloud

18. Design a REST API for a library's books and loans.

"GET /books, GET /books/{id}, POST /books, PUT /books/{id}, DELETE /books/{id}; loans as POST /loans with book and member IDs, GET /members/{id}/loans, PATCH /loans/{id} to return a book. Proper status codes, pagination on lists, validation errors as 400 with details, and authentication with a bearer token. I would version the API from day one."

19. What happens when a browser requests a page from your application?

"DNS, TCP and TLS handshakes, HTTP request through a load balancer to the app server, server-side logic and database calls, response with HTML or JSON, then the browser fetches assets and renders. I would mention caching headers and a CDN for static assets because Accenture panels often follow up on performance."

20. Explain how you use Git in a team.

"Feature branches from main or develop, small pull requests with review, rebase or merge from main before opening the PR, and a CI check that runs tests on every PR. Conflicts are resolved locally by the branch owner. Tags for releases and a hotfix branch process for production."

21. What is CI/CD, and what did your pipeline do?

"Continuous integration builds and tests every commit; continuous delivery deploys automatically to an environment after tests pass. Our Jenkins pipeline ran unit tests, static analysis, built a Docker image, pushed it to a registry and deployed to a dev environment; production needed a manual approval."

22. Which cloud services have you used, and what would you use to host a small web app?

Name real usage. "AWS EC2 for a VM, S3 for uploads, RDS for Postgres. For a small app today I would use a managed platform, App Service on Azure or Elastic Beanstalk on AWS, a managed database, and object storage for files, with secrets in a vault, not in config files."

23. What are Docker containers, and why use them?

"A container packages the application with its dependencies so it runs the same on a laptop, in CI and in production. Lighter than a VM because it shares the host kernel. I wrote a Dockerfile for the API and a Compose file for the API plus database, which fixed the 'works on my machine' problems in our team."

Delivery scenarios and HR

24. A user story is estimated at three days and on day two you realise it will take six. What do you do?

"Tell the scrum master and product owner on day two, not day six, with the reason and options: deliver a smaller slice this sprint, or move the story. I would also note what I missed in estimation so the next one is better. Accenture is a client-delivery business; hiding a slip is the worst outcome."

25. How do you handle a client who keeps changing requirements?

"Write the requirement down and get it confirmed, show the impact of each change on scope and dates, and route changes through the product owner or change process rather than absorbing them silently. Frequent changes usually mean the client is unsure; a quick prototype or demo settles it faster than arguing."

26. Explain your project's architecture and one decision you would change.

Layers, data flow, hosting, your part, and one honest regret. "I would move email sending out of the request path into a queue; it made the API slow whenever the mail server was slow."

27. Tell me about yourself.

Sixty to ninety seconds: education or current role, primary skill with proof, one project result, why Accenture Custom Software Engineering. Practise aloud; the tell me about yourself guide has templates.

28. Are you open to relocation, rotational shifts and working on a different technology?

"Yes to relocation across India, yes to shifts when a client needs them, and yes to a different stack; I moved from PHP to Java in my last project in a month." Accenture allocates by client need; a hard no here ends most fresher applications.

29. Why Accenture, and why not a product company?

"The variety of clients and technologies in the first few years, structured training, and the chance to work with global teams. I want breadth before depth, and Accenture's Custom Software Engineering work gives that."

30. What is your notice period and expected CTC?

Have facts ready: notice period, negotiability, buyout. For CTC, ask the recruiter for the level and band first, then give a range. Do not quote a number from a forum.

Mistakes that get candidates rejected

• Not knowing what Custom Software Engineering means and preparing for the wrong role.
• Weak primary skill. The JD names it; the interview goes deep on it.
• Definitions without project examples in OOP and SQL answers.
• No Git, CI/CD or cloud vocabulary for a 2026 opening.
• A conditional yes on relocation or shifts in HR.
• Talking about "we" for the whole project and never "I".

How to practise the Accenture Custom Software Engineer round

The technical round is a conversation about your primary skill and your project with coding in the middle. The candidates who pass have said their answers aloud, been interrupted, and fixed the weak ones.

In MockMate Practice, attach your resume and paste the Accenture job description with its primary skill, choose a technical round and run an adaptive session in the browser. The interviewer persona probes your project, asks language and SQL questions with follow-ups, and sets a coding problem you solve in the built-in editor. The report shows each answer, response timing, coaching evidence and recurring weaknesses, with a recommended next practice. Eligible accounts get three free Practice starts of up to ten minutes each as of September 2026. Use MockMate Live assistance only in interviews or meetings where the organisation or interviewer permits it; Accenture selection rounds do not, so stay with Practice for this interview.

A one-week plan: two problems a day from arrays, strings and stacks, spoken through first; one day on SQL joins and window functions by hand; one day rewriting your project as architecture, decision and regret; a Practice round, a report read, and a second round on the weakest group. For the manager conversation see the Accenture managerial round page, and for the wider process the Accenture hub.