Explore: Interview experience · HR interview questions · Overview · Placement papers · Preparation guide
Salesforce’s loop runs a recruiter screen, an OA, two technical rounds skewed toward practical platform design, and a values/hiring-manager round built on V2MOM and Ohana.
Interview experience
Salesforce interview experiences covering Recruiter / HR phone screen, Online Assessment (HackerRank), Technical Round 1, Technical Round 2, +2 more. Round-wise notes, common questions, and what candidates report clears each stage.
HR interview questions
Salesforce HR interview questions with sample answer frameworks - motivation, relocation, STAR stories, and culture-fit prompts.
| Round | Duration | What it tests |
|---|---|---|
| Recruiter / HR phone screen | 20-30 min | Background, motivation, comp expectations, logistics |
| Online Assessment (HackerRank) | 60-90 min | 2-3 DSA problems, easy to medium-hard |
| Technical Round 1 | 45-60 min | DSA coding + detailed project/resume discussion |
| Technical Round 2 | 45-60 min | Practical system/platform design (rate limiter, scheduler, multi-tenant data model) |
| Hiring Manager / values round | 30-45 min | Ohana culture fit, V2MOM-style goal framing, team fit |
| HR / offer discussion | 20-30 min | Compensation, location, closing logistics |
These answers are compiled from public candidate reports and off-campus job listings, not from an official Salesforce notice, and drives differ by campus and year. The instruction screen of your own test is the final word.
| Question | Answer |
|---|---|
| Which platform is it on? | HackerRank. AMTS candidate reports from 2024 to 2026 consistently name it. |
| How many questions and how long? | Usually three coding problems in 90 minutes. Some 2025-2026 reports describe 90-100 minutes, and older campus drives had two coding rounds. |
| How hard are the problems? | Reports conflict. A February 2026 report describes three medium-to-hard problems on graphs, greedy and dynamic programming; a 2026 Hyderabad report calls the three questions easy. A May 2024 Bengaluru test had graph BFS, a variable-size sliding window and 2D DP. |
| Does it have negative marking? | No. It is a coding-only test scored on hidden test cases, and no candidate report mentions MCQ penalties. |
| Is there a sectional cutoff? | There are no sections, only coding problems. Candidate reports say clearing roughly two of the three problems usually gets a shortlist, but no score is published. |
| How much is the cutoff? | No test cutoff is published. AMTS job listings set eligibility at 7.0+ CGPA with no active backlogs for 2025 and 2027 batches; some 2026-batch listings say 7.5+, so check your own listing. |
| Which languages can I use? | Java, Python and C++ are the languages candidates report using; HackerRank supports others, but check the test settings. |
| Is there an off-campus route? | Yes. Salesforce posts Software Engineering AMTS openings for Bengaluru and Hyderabad on its careers site for the graduating batch, under its Futureforce university-recruiting programme. |
These themes come from candidate reports on Glassdoor and LeetCode Discuss for Salesforce’s Hyderabad and Bengaluru offices. They show the level and style, not a fixed list.
| Role | Reported questions and topics |
|---|---|
| AMTS (new-grad software engineer) | A June 2024 AMTS loop had an OA with Word Break, then a 60-minute technical round on median from a data stream and the shortest path between two nodes in a binary tree, then a 30-minute managerial round on the internship, React’s virtual DOM, strengths and weaknesses. Late-2025 Hyderabad reports describe one technical round on two medium DSA questions, projects, OS and OOP, then HR. |
| AMTS intern (Futureforce) | Same HackerRank OA, then one or two rounds on DSA and resume projects. A 2026 Hyderabad report lists basic low-level design questions plus binary search and merging two sorted linked lists. |
| MTS / SMTS (experienced) | Coding plus a design round. Candidates report a second round focused on system design (a 2026 Noida candidate was rejected there) and platform-flavoured problems such as a rate limiter, job scheduler or multi-tenant data model. |
| Behavioural / managerial (all roles) | Handling conflict and difficult situations, a goal described in V2MOM shape, and questions tied to Trust and Customer Success. |
Asked across every role: a deep walk through one project (your contribution, the hardest bug, what you would change), OOP basics, and why Salesforce.
A short call to confirm background, work authorization, location, and comp expectations before any coding starts. It’s a filter, not a technical bar - be direct about your timeline and current CTC/expected CTC rather than vague.
Common questions
Full behavioural frameworks are on the Salesforce HR interview questions page.
A 60-90 minute HackerRank test with 2-3 coding problems ranging from a warm-up array/string question to a medium-hard one, plus sometimes a parsing or matrix-traversal problem. 2025-2026 India AMTS reports also mention 90-100 minute OAs with three problems, including bit-manipulation / math (reduce an integer using ± powers of two), range updates on arrays, and tree or graph tasks (diameter endpoints, organizational hierarchy trees). Candidates who clear roughly two of the three problems typically get shortlisted for technical rounds - full optimality on every problem is less common than the bar at some other product companies.
Common questions
See how real candidates handled this stage on the Salesforce interview experience page.
A 45-60 minute round that opens with a walkthrough of your resume or a past project - architecture, your specific contribution, and the hardest bug you hit - before moving into 1-2 coding problems. Interviewers weigh how clearly you can explain a real system you built as much as the coding itself.
Common questions
Round-by-round breakdowns are on the Salesforce interview experience page.
This is where Salesforce’s process diverges most from a generic Big Tech loop: rather than a textbook distributed-systems design, candidates report being asked to design something with a clear platform flavor - a rate limiter, a notification system, a background job scheduler, or a simplified multi-tenant data model, since multi-tenancy is core to how Salesforce’s own CRM platform works.
Common questions
A 30-45 minute conversation that functions as Salesforce’s version of a “culture fit” round, built around its stated values (Trust, Customer Success, Innovation, Equality, Sustainability) and the broader Ohana (“family”) framing of employees, customers, and partners as one extended community. Some interviewers ask you to describe a project or goal using Salesforce’s internal V2MOM structure rather than plain STAR.
Common questions
Sample answer frameworks for each of these are on the Salesforce HR interview questions page.
V2MOM (Vision, Values, Methods, Obstacles, Metrics) is the planning framework Salesforce has used company-wide for over two decades - every team’s goals, and the CEO’s own, are written in this format and shared internally. It’s genuinely distinctive versus a typical STAR-only behavioural round: some hiring managers ask candidates to restate a past project or goal in V2MOM shape (what you wanted, why it mattered, how you’d get there, what stood in the way, and how you’d measure success) instead of a plain Situation-Task-Action-Result story. Rehearsing one story this way - not just as STAR - is a small, specific edge for the values round.
Many people searching for “Salesforce interview questions” are preparing for Salesforce developer or admin jobs at partner and services firms, where the work is building on the Salesforce platform: Apex, SOQL, triggers, governor limits, Lightning Web Components, flows and sharing rules. Those interviews test platform knowledge and certifications, not a DSA-heavy coding loop.
This page covers the other job: engineering roles at Salesforce itself (AMTS, MTS, SMTS) in Hyderabad and Bengaluru, where you build the platform. That loop is a HackerRank coding test, DSA and design rounds, and a values round. You do not need Apex or Salesforce certifications for it, though knowing that the platform is multi-tenant helps in the design round.
Maintain four boundaries - top, bottom, left, right - and loop while top is within bottom and left is within right. Each iteration traverses left to right along the top row then increments top, top to bottom down the right column then decrements right, right to left along the bottom row then decrements bottom, and bottom to top up the left column then increments left. The two traversals after the first need a guard check that the boundaries have not crossed, otherwise a single remaining row or column is printed twice. Time is O(n times m) since each cell is visited once, and space is O(1) beyond the output.
Push every opening bracket onto a stack. On a closing bracket, the stack must be non-empty and its top must be the matching opener, otherwise return false immediately; pop on a match. At the end the string is balanced only if the stack is empty, which catches unclosed openers. Use a map from closing to opening bracket so all three bracket types are handled uniformly. Time is O(n) and space is O(n) in the worst case of all openers. The same shape solves the HTML-tag-ordering variant Salesforce asks: parse each tag, push opening tags, and require the matching name on the top of the stack for each closing tag.
Use a modified binary search. At each step, at least one half of the array is still sorted: compare the value at low with the value at mid to find out which. If the left half is sorted and the target lies between the values at low and mid, recurse left, otherwise recurse right - and symmetrically when the right half is sorted. That keeps the search O(log n) time and O(1) space without ever finding the pivot separately. With duplicates the check at low equal to mid becomes ambiguous, and you must shrink low by one, which degrades the worst case to O(n).
The costly-push version keeps the stack order in q1 at all times: to push, enqueue the new element into the empty q2, dequeue everything from q1 into q2, then swap the names. Push is O(n), while pop and top are O(1) because the newest element is already at the front. The costly-pop version simply enqueues into q1 on push, and on pop moves all but the last element into q2 before dequeuing that last one - O(1) push, O(n) pop. Choose based on the expected operation mix; interviewers usually want you to state both and justify the pick.
BFS explores level by level using a queue and is the right choice when you need the shortest path in an unweighted graph, the minimum number of moves, or level-order structure, because the first time it reaches a node it has done so in the fewest edges. DFS uses a stack or recursion and is better for exhaustively exploring paths - cycle detection, topological sort, connected components, and backtracking problems - and it usually uses less memory on wide graphs, since BFS can hold an entire level in the queue. Both are O(V + E) time; the space difference is O(width) for BFS versus O(depth) for DFS.
There are three common patterns. A separate database per tenant gives the strongest isolation and easy per-tenant restore, but scales poorly past a few hundred tenants. A shared database with a schema per tenant is a middle ground. A fully shared schema with a tenant_id column on every table - the model Salesforce’s own platform uses - scales to very large tenant counts and is cheapest to operate, but isolation now depends entirely on discipline. In that model, tenant_id must be the leading column of every primary key and index so queries prune to one tenant’s rows, and the filter should be enforced centrally through row-level security or a query layer rather than trusted to every hand-written query.
The client redirects the user to the authorization server, which authenticates them and redirects back with a short-lived authorization code. The client then exchanges that code, plus its client secret, for an access token over a back-channel call the browser never sees - which is why the code, not the token, travels through the redirect. Public clients add PKCE, sending a code challenge up front and the verifier at exchange time, so an intercepted code is useless. A JWT is a token format, not a protocol: a base64url header, claims payload, and signature, self-contained so the resource server can validate it with a public key without calling the auth server. That statelessness is also its weakness - a JWT cannot be revoked before expiry, so keep access tokens short-lived and use refresh tokens for longevity.
Put a queue between the producing service and delivery so a spike in events never blocks the write path. Events land in a topic; workers fan out per recipient, look up preferences and channel (email, SMS, push, in-app), render from a template, and hand off to the channel provider. Deduplicate on an event ID so a retried producer does not double-send, and store delivery state per notification so retries are safe. Respect user preferences and quiet hours at fan-out time, not at send time, and batch or digest high-frequency events rather than sending each one. Provider failures need exponential backoff with a dead-letter queue, and separate queues per priority keep a backlog of marketing sends from delaying a password reset.
Start from requirements: limit per user, per tenant, or per API key; window type (fixed, sliding, token bucket). A token bucket allows steady traffic with bursts; a fixed window is simpler but can allow double bursts at boundaries. Store counters in a low-latency store (Redis or in-memory sharded counters with sync) keyed by client ID plus window bucket. On each request, decrement tokens or increment count; return HTTP 429 with Retry-After when exceeded. For Salesforce-style multi-tenant load, isolate counters per tenant so one noisy customer cannot exhaust a shared bucket - reported interview follow-ups often probe exactly that failure mode.
Separate submission from execution: producers enqueue jobs with type, payload, priority, and earliest run time. Workers poll or subscribe to a priority queue; claim jobs with atomic leases so two workers never run the same job. Persist state (pending, running, succeeded, failed, dead-letter) for observability and safe retries with idempotency keys. Cap concurrent jobs per tenant and globally to protect shared databases - analogous to governor-limit thinking on Salesforce’s own platform. Schedule delayed jobs with a time-ordered structure or delayed queue; on failure, exponential backoff before dead-lettering for manual replay.
FutureForce is Salesforce’s university hiring program for interns and new graduates. In India, engineering hiring concentrates in Hyderabad (with some Bangalore roles) for platform, Sales Cloud, Service Cloud, MuleSoft, and AI-related teams. Loops still include HackerRank OAs and multiple technical touchpoints, but cohort timing and eligibility (CGPA floors often around 7.0-7.5 in drive mails) vary by year - always treat the official careers posting and your invite as source of truth, not third-party cutoff rumors.
No. The Salesforce OA is a HackerRank coding test, usually three problems in 90 minutes, scored on hidden test cases, and no candidate report mentions any penalty. Candidates report that solving about two of the three problems fully usually gets a shortlist.
No test cutoff is published; the fixed filter is eligibility. AMTS job listings for the 2025 and 2027 batches ask for 7.0+ CGPA with no active backlogs in CS, electrical, electronics or related branches, while some 2026-batch listings say 7.5+. Check the CGPA line on your own listing.
Salesforce’s loop usually runs: 1. Recruiter/HR phone screen (20-30 min) - background, motivation, comp expectations. 2. Online Assessment (60-90 min, HackerRank) - 2-3 DSA problems at easy-to-medium-hard difficulty. 3. Technical Round 1 (45-60 min) - DSA coding plus a deep dive into your resume/projects. 4. Technical Round 2 (45-60 min) - practical system/platform design (rate limiters, notification systems, job schedulers, multi-tenant data models) rather than pure LeetCode. 5. Hiring Manager / values round (30-45 min) - Ohana culture fit and V2MOM-style goal framing. End-to-end timeline is roughly 3-5 weeks, sometimes up to 6 for senior or specialized roles.
Coding rounds lean on arrays/strings, matrix traversal (e.g. spiral order), stack/queue implementations, graph BFS/DFS, and parsing-style problems (balanced parentheses, checking if HTML tags are correctly ordered) at medium difficulty. The second technical round often shifts to practical platform design - a rate limiter, notification system, job scheduler, or multi-tenant data model - reflecting Salesforce’s own CRM/cloud platform rather than textbook distributed-systems design. Values rounds sometimes ask you to frame a project or goal using Salesforce’s own V2MOM structure.
Typically 4-5 touchpoints: a recruiter/HR screen, an online assessment, 2 technical rounds (coding, then practical system/platform design), and a hiring-manager/values round - sometimes with a separate HR/offer call. Composition varies by team and level, so treat this as a template, not a fixed script.
Drill medium-difficulty DSA (arrays, graphs, stacks/queues, matrix traversal) alongside a few practical design problems (rate limiter, notification system, job scheduler), be ready to describe a project’s goal using Salesforce’s V2MOM framework (Vision, Values, Methods, Obstacles, Metrics), and prepare STAR stories that show trust and customer-first thinking - Salesforce’s own stated top two values.
V2MOM (Vision, Values, Methods, Obstacles, Metrics) is the internal planning framework Salesforce uses company-wide, from the CEO’s own annual goals down to individual teams. Interviewers - especially in the hiring-manager/values round - sometimes ask you to describe a past project or goal using this structure, so it’s worth rehearsing one story in V2MOM shape rather than a plain STAR format.
Yes, through FutureForce, Salesforce’s global university program that runs real campus and off-campus drives for interns and new grads, concentrated around its Hyderabad engineering hub (with some Bangalore roles too). That puts Salesforce closer to Amazon’s campus-hiring model than to companies like Netflix that skip campus drives almost entirely - though volumes are smaller than mass recruiters like TCS or Infosys.
Ohana is Hawaiian for family, and Salesforce uses it to describe its extended community of employees, customers, and partners, built around stated values of Trust, Customer Success, Innovation, Equality, and Sustainability - in that order. In interviews it shows up as questions about building trust with a team or customer, prioritizing customer success over a faster shortcut, and working with people whose styles differ from yours.
Product engineering loops usually combine OA or coding screens, technical depth, and a values-aware behavioural close. Interviewers score:
Soft fail modes: CRM buzzwords without a project, or behavioural answers that skip the customer impact.
Public guides treat the values/behavioural stage as heavily weighted, not optional. Map STAR stories to Trust, Customer Success, Innovation, Equality, and Sustainability with a measurable Result. Jumping straight to code in live rounds without stating approach and complexity is also a common soft fail in Salesforce coding interviews.
Good fit: You triage HackerRank problems in three minutes, write custom tests before submit, can defend one project with metrics, and have real STAR stories for Trust and Customer Success - not slogan bingo. Internships help; a measurable personal project beats a vague CRM buzzword resume.
Harder path: You confuse SWE OA with Admin certification prep; you over-invest in the hardest graph while an easy array case fails; your Why Salesforce is “brand + CTC.” Values rounds are real interviews.
Use this as a default four-week loop before a Salesforce drive. Adjust to your invite timeline.
10-day crash version: Days 1-4 timed DSA, days 5-7 resume projects + CS fundamentals, days 8-9 mock interviews out loud, day 10 HR STAR stories on the HR page.
Candidate-reported fresher/early-career packages for Salesforce often cite ₹25-40 LPA (AMTS / fresher SWE) · Languages Allowed: Java, Python, C++. Treat forum numbers as ranges, not guarantees - confirm the written offer letter for your team and location.
Interview process details and sample questions on this site are editorially organized from public candidate reports and hiring write-ups, with round structure, prep notes, and company-specific framing added by our team. For round-by-round stories, see the Salesforce interview experience page. For OA papers and coding drills, use Salesforce on placementpapers.app. OA quick answers and role-wise questions draw on Glassdoor candidate reviews for Salesforce Hyderabad, Bengaluru and Noida (2024-2026), a June 2024 AMTS interview experience on LeetCode Discuss, and Salesforce Software Engineering AMTS listings reproduced on Peerlist and fresher job sites.
Guides often searched alongside this page: Oracle interview questions · SAP interview questions · Microsoft interview questions · ServiceNow interview questions · Adobe interview questions · Google interview questions.
