Decagon Software Engineer interview questions

26 Software Engineers have shared notes from their Decagon interviews, with the questions they were asked in each round. That is 96% of the 27 Decagon interview notes on Screna AI.

26 Software Engineer interview notes

Decagon·Software Engineer·Technical Phone Screen

Jun 2026
45-minute phone screen at Decagon for a software engineer role, one coding question the whole time. Made two bugs even though I'd seen this type of problem before, but caught them during testing and the interviewer let it slide.
  • Design a CSAT (Customer Satisfaction) tracker class that accepts a sliding time window. It should support: initializing with a window size, recording a conversation ID with a score (1-5) and a strictly increasing timestamp, and returning the average score of all entries within the current active window rounded to 2 decimal places.
  • Follow-up: add an update method that takes a conversation ID and a new score. If the conversation is still within the active window, update its score. If it has expired, do nothing. The original timestamp and ordering should remain unchanged.
  • Second follow-up: add a percentile method that, given an integer p, returns the score at that percentile across all conversations currently in the active window.

“I'd done something like this before so I felt confident going in.”

View Post

Decagon·Software Engineer·Onsite - System Design / Architecture

Jun 2026
System design round at Decagon for a software engineer role, focused entirely on building an AI Gateway service. The scope was narrower than a typical distributed systems question since scaling was explicitly off the table, but the depth required on routing, rate limiting, and observability made up for it.
  • Design an AI Gateway service that sits between internal application services and external model providers like OpenAI, Anthropic, and Google.
  • How would you handle rate limiting across tenants, routes, and individual providers in this gateway?
  • Walk me through how you'd instrument this gateway for metrics and logging, and how that data flows to dashboards and alerting.
  • What other concerns would you address in this gateway beyond routing and rate limiting?

“I went straight to the API layer and request flow, which felt right, but I underweighted the routing logic early on and had to backtrack.”

View Post

Decagon·Software Engineer·Onsite - System Design / Architecture

Jun 2026
System design round at Decagon for a software engineer role. The whole thing was one big question about building an AI gateway and they really wanted you to go deep on every layer of it, not just sketch a box diagram and call it done.
  • Design an AI Gateway service that sits between application clients and multiple upstream LLM providers, exposing a unified API. Cover failover, fast response, cost tracking, accuracy guarantees, rate limiting, and observability. Walk through your data model, request flow, and key failure modes.
  • How would you handle failover and circuit breaking when an upstream LLM provider starts returning errors or timing out?
  • What strategies would you use to reduce latency and improve response speed, including caching and parallel provider calls?
  • How would you design the cost tracking and budget enforcement system, including per-tenant attribution and a cost dashboard?
  • How would you ensure response accuracy and quality, including schema validation, consistency checks across providers, and detecting when a provider's output quality has degraded?

“This is a sprawling question and I underestimated just how many subsystems they expected you to address.”

View Post

Decagon·Software Engineer·Technical Phone Screen

Jun 2026
Interviewed for a software engineer role at Decagon and got a classic stack-based CPU scheduling problem. Pretty straightforward if you've seen this type before, but the edge cases around nested calls can trip you up if you're not careful.
  • Given a list of log strings representing function start and end events on a single-threaded CPU, compute the exclusive execution time for each function (i.e., time spent running but not inside a nested call).

“I knew immediately this was a stack problem but fumbled the timestamp arithmetic for a bit.”

View Post

Decagon·Software Engineer·Online Assessment (OA)

Jun 2026
Decagon software engineer interview, part of what looks like a multi-part coding assessment. This section was all about implementing undo/redo on a shopping cart, which sounds manageable until you actually have to think through the edge cases.
  • Implement undo and redo functionality for a shopping cart system, where performing a new operation resets the redo history, and get_total_discount() always reflects the current state.

“The basic stack approach for undo/redo clicked pretty fast.”

View Post

Decagon·Software Engineer·Technical Phone Screen

Jun 2026
Decagon SWE interview had me building out a rolling-window CSAT tracker, which started reasonable enough and then they threw in an update operation that forced me to think harder about the data structure underneath.
  • Design a rolling-window CSAT tracker that supports recording scores, computing averages over a time window, and updating a score in place by conversation ID. How do you index by conversation ID to make the update operation O(1)?

“The record and get_average parts came together fine, deque with lazy eviction is pretty standard.”

View Post

Decagon·Software Engineer·Technical Phone Screen

Jun 2026
Interviewed for a software engineer role at Decagon and got a stack-based simulation problem. Pretty straightforward round, nothing too surprising.
  • Given n functions running on a single-threaded CPU and a list of logs in the format 'function_id:start|end:timestamp', compute the exclusive execution time for each function (meaning time spent in that function itself, not in any functions it called).

“Stack simulation problem.”

View Post

Decagon·Software Engineer·Technical Phone Screen

May 2026
Decagon SWE interview had a design-your-own-data-structure problem centered on a rolling window CSAT tracker. Pretty focused session, one meaty question with a lot of follow-up on complexity tradeoffs.
  • Design and implement a CSATTracker class that records customer satisfaction scores and returns a rolling average over a configurable time window. Discuss your choice of data structure and the time and space complexity of each operation.

“I jumped straight to a deque sorted by timestamp, which was the right call, but I fumbled explaining the eviction logic.”

View Post

Decagon·Software Engineer·Technical Phone Screen

May 2026
Decagon coding interview, one algorithmic question on function call log parsing. Pretty focused session, no fluff.
  • Given logs of n function calls on a single-threaded CPU, each in the format 'function_id:start|end:timestamp', compute the exclusive time for each function (time spent in the function itself, excluding any time spent inside nested calls). Return an array of length n.

“Stack-based solution is the right move here.”

View Post

Decagon·Software Engineer·Onsite - System Design / Architecture

May 2026
System design round at Decagon for a software engineer role. One big question, very deep, lots of follow-ups baked into it. Left feeling like I probably undercooked the observability side.
  • Design an AI gateway that sits between internal services and multiple LLM providers, handling routing, fallback logic, and structured observability for business teams.
  • Which specific signals should drive a provider failover, and how do you prevent the system from flapping between providers?
  • How would you design the log data model to support per-team usage reporting, token counts, cost tracking, and latency breakdowns?
  • Should the fallback policy be static or adaptive, and what are the tradeoffs?

“This question is basically three questions stapled together.”

View Post