Freshworks Interview Questions 2026

Tip: Freshworks: Freshdesk (customer support), Freshsales (CRM), Freshservice (IT service management), Freshchat (live chat). Engineering focus: full-stack web, search and filtering at scale, real-time messaging, integrations ecosystem (1000+ third-party connectors). Show product awareness in your motivation answer.

Tip: Core entities: Ticket, Agent, Customer, Team, SLA. Key features: email-to-ticket ingestion (webhooks from mailbox), assignment rules engine (round-robin, skill-based), SLA timer service (cron-based), real-time agent collaboration (WebSockets), full-text search (Elasticsearch on ticket body + metadata).

Tip: Full-text search: find documents containing query terms with relevance ranking. Elasticsearch: inverted index maps terms to document IDs. Tokenisation, stemming, stopword removal applied at index time. Querying: BM25 scoring by default. Fuzzy search handles typos. Freshdesk uses Elasticsearch for ticket search across billions of records.

Tip: State: tokens (float), last_refill (timestamp). On request: compute elapsed time, add tokens = elapsed × rate, cap at max_tokens, check if tokens >= 1: if yes consume 1 and allow; else reject. Thread-safe: use mutex or atomic operations. Freshworks API has per-account rate limits on their platform.

Tip: Rails: convention-over-configuration web framework in Ruby. Freshdesk was originally built in Rails. Know: MVC pattern in Rails, ActiveRecord ORM, RESTful routing conventions, asset pipeline. Freshworks is migrating parts to React frontends + Go/Java services, but Rails is still the core — knowing it is a differentiator.

Tip: Connection pooling: maintain a pool of reusable DB connections rather than creating a new connection per request (expensive: TCP handshake + auth). Tools: PgBouncer (Postgres), HikariCP (Java). Freshworks serves thousands of simultaneous tenants — without pooling, each concurrent request would open a new DB connection, exhausting DB limits.

Tip: Freshworks values engineering excellence. Examples: automated a repetitive test setup, built a CLI tool for common deployment tasks, introduced linting rules that prevented a class of bugs. Frame it as: problem identified → effort estimated → impact measured (saved X hours/week for the team).

Tip: Options: (1) WebSockets — persistent bi-directional connection, ideal for chat. (2) Server-Sent Events (SSE) — one-way server push, simpler, works over HTTP. (3) Long polling — fallback for environments blocking WebSocket. Freshchat uses WebSockets + fallback. For scale: pub-sub via Redis/Kafka for fan-out to multiple connected clients.

Tip: N+1: fetching 1 parent record then N separate queries for each child — instead of 1 JOIN. In Rails: add `.includes(:association)` for eager loading. In Django: `.select_related()` for FK, `.prefetch_related()` for M2M. In JOOQ/Hibernate: fetch joins. Always profile with query count logging in development.

Tip: Freshworks Marketplace: webhook + REST API-based integrations built by partners. Core integrations (Slack, Jira, Salesforce) maintained by Freshworks engineering. Platform team provides SDK for building apps. Interview relevance: design a webhook integration with retry logic, signature verification (HMAC), and event idempotency.