The SIG coding assessment is the timed online coding test that applicants for Susquehanna International Group (SIG) developer roles report taking early in the process. SIG's job postings for some technology roles list a technical screen or technical assessment as the first step after you apply, but SIG does not publish the format. According to candidate reports, it runs on CodeSignal with 4 questions in about 70 minutes, starting easy and getting harder, and each question is scored when you submit it. Quant research applicants report a math and probability test instead. Recruiter, technical and final interviews follow.
From SIG's own job postings and CodeSignal's help center, checked November 3, 2026. SIG does not publish its test format, so most details here come from candidate reports and say so. Your invitation email wins if it says something different.
| What SIG confirms | Some technology postings list a "Technical Screen", "Technical Assessment" or "Online Assessment" right after you apply |
|---|---|
| Platform | CodeSignal, according to candidate reports from 2024 to 2026; SIG does not name it |
| Questions and time | Candidates report 4 questions in about 70 minutes; some say it is CodeSignal's General Coding Assessment (GCA) |
| Difficulty | Candidates report 1-2 easier questions first, then harder ones |
| Invite | Candidates report an emailed link; one got it right after applying, with a deadline more than a month out |
| Languages | SIG's developer postings say it uses C#, C++ and Python mostly; the CodeSignal GCA lists 38 languages |
| Quant research roles | Candidates report a timed logic, combinatorics and probability test instead (one reports 17 questions in 1 hour) |
| After the test | Recruiter interview, technical interview(s), final interview, offer (SIG postings) |
| Integrity checks | On CodeSignal, employers see a Suspicion Score for each question |
| Proctoring | Not published by SIG; CodeSignal shows before you start whether camera, microphone and screen recording are required |
| Retakes | Not published by SIG; CodeSignal limits retakes by test type and needs a new invite |
| Free practice | CodeSignal's practice assessments; companies cannot see your activity |
SIG does not publish its questions. These are the types candidates describe, written in our own words, with original tips.
A short problem with clear rules: transform a list of numbers or strings and return the result. Candidates report one or two of these at the start.
Tip: Solve these fast and move on. The time you save is what the last two questions need.
Read structured text or a stream of events and apply rules step by step. One candidate reports a parsing question took far longer than expected.
Tip: Split the input handling from the rules. Most lost time here is fighting the input format, not the logic.
Harder questions where a brute force times out on large hidden tests, so the right hash map, heap, sort or graph search matters.
Tip: Submit a correct slow version first so you bank partial points, then optimize.
Build a class with a few methods that must stay correct across many calls, a style quant developer candidates describe.
Tip: Keep state in one place and test each method with a tiny example before you submit.
Quant research candidates report a separate timed test of logic, counting and probability puzzles rather than code.
Tip: Drill expected value and counting with a timer. Speed under pressure is what these tests measure.
SIG does not publish a passing score or how it weighs the test. Candidates report that each question carries its own points and is scored when you submit it. Candidates who move on go to recruiter and technical interviews.
If your invite is for CodeSignal's General Coding Assessment, CodeSignal's GCA format applies: 4 questions in 70 minutes in one sitting, any order, and your best submission counts. Submitting runs hidden test cases, so partial credit is normal. Certified results are scored from 200 to 600, you can see your score in your CodeSignal account, and CodeSignal publishes no passing score.
Take CodeSignal's free practice assessments and do at least one set of 4 questions in 70 minutes, the length candidates report for SIG. Companies cannot see your practice activity.
Clear two easier questions quickly, then spend most of your time on the harder two. Submit after each one so your points are saved.
Practice turning messy input into clean data, simulating rules step by step, and picking a data structure that survives large hidden tests.
Use the language you are fastest in. One candidate lost time when a C++ library feature did not build, so try the standard library pieces you rely on in CodeSignal's editor first.
SIG does not say whether its coding test is proctored or what it tracks. Your invite and the start screen tell you.
If it runs on CodeSignal, as candidates report, CodeSignal shows whether proctoring is required before you begin. Proctored CodeSignal tests record your camera, microphone and screen for the whole session and check a government photo ID. Employers also see a Suspicion Score for each question, built from signals like code similarity, pasting and session replay.
After the assessment, SIG tests the same skills live: its postings list technical and final interviews after the assessment, and candidates describe live coding and design rounds.
Lists of reported SIG questions circulate, but SIG does not publish its questions, and on CodeSignal your score depends on hidden test cases for the exact problems you get, while the Suspicion Score looks at code similarity. Coding Copilot reads the problem actually on your screen and writes a solution for that one.
OphyAI works on the real test: Coding Copilot in the OphyAI desktop app snaps the problem with one hotkey and shows an approach, complexity notes and working code in an overlay that screen sharing and recording leave out. In SIG's live technical interviews, Space or Cmd/Ctrl + Shift + L transcribes the question your interviewer asks out loud. It does not hide from a webcam or from CodeSignal's own checks.
SIG does not say what it checks. On CodeSignal, proctored tests record your camera, microphone and screen and check a photo ID, and employers see a per-question Suspicion Score built from code similarity, pasting and session replay. The OphyAI window is hidden from screen sharing and recording, but not from a webcam, a person watching, or the platform's own keystroke, paste and tab tracking.
Compare every platform: Can online assessments detect AI?
Coding Copilot in the OphyAI desktop app reads the coding problem on your screen with one hotkey and returns an approach, complexity notes and working code in your language.
Use OphyAI Coding Copilot on your real SIG coding assessment, not just to practice. Press Cmd/Ctrl + Shift + A and it snaps the problem window, even when other windows cover it, or press Cmd/Ctrl + Shift + S to drag a box around one part of the screen. The answer shows as a card in a floating overlay that stays on top of a full-screen browser and opens without taking the keyboard from your editor: how it read the problem, the approach, code with line numbers, a dry run with edge cases and the Big O. Follow-ups remember the problem, so "now handle an empty input" updates the same code.
If your invite is for a proctored CodeSignal session, it records your camera, microphone and screen. In SIG's live technical interviews afterwards, Space or Cmd/Ctrl + Shift + L transcribes the question your interviewer asks out loud.
Cmd/Ctrl + Shift + A snaps the problem window, even when other windows cover it, and Cmd/Ctrl + Shift + S lets you drag a box around part of the screen. In a live interview, Space or Cmd/Ctrl + Shift + L transcribes the question the interviewer asks out loud.
Answers show as a card in a floating overlay that macOS and Windows leave out of screen sharing and recording. It stays on top of a full-screen editor or browser and opens without taking the keyboard from your editor.
Each answer card shows how OphyAI read the problem, what to say first, the code with line numbers, a dry run with edge cases, and the Big O. Follow-ups remember the problem, so "now handle duplicates" updates the same code.
Free download. Coding Interview is on the Premium plan ($59 a month, 500 credits) and uses 15 credits per session. See everything it does
Hiding a window from screen share does not hide it from a webcam, a person watching, keystroke and paste history, tab or focus tracking, or proctoring software.
SIG does not publish a time. Candidates most often report about 70 minutes for 4 questions. Your invite shows the real limit.
Candidate reports from 2024 to 2026 say CodeSignal, and some name its General Coding Assessment. SIG does not name the platform, so check your invite email.
Candidates describe one or two easier questions followed by harder ones, and one reports a parsing question that took far longer than planned. A full timed practice set helps more than any single topic.
Software candidates describe coding only. Quant research candidates report a separate timed test of logic, combinatorics and probability; one reports 17 questions in 1 hour.
SIG does not say. If it is a proctored CodeSignal assessment, you share your camera, microphone and screen and show a photo ID, and CodeSignal tells you before you start.
The one you solve fastest in. SIG's developer postings say it mostly uses C#, C++ and Python, and the CodeSignal GCA lists 38 languages. Test any library features you rely on in CodeSignal's practice editor first.
SIG does not publish a retake policy. On CodeSignal, retakes are limited by test type and always need a new invite, and for certified tests you can share an eligible earlier result instead if the company accepts it.
SIG's postings list a recruiter interview, technical interviews and a final interview before an offer. Candidates describe live coding and design rounds for software roles, and probability, statistics and algorithms questions for quant research roles.
Read on November 3, 2026. OphyAI is not affiliated with Susquehanna International Group.
Download OphyAI Desktop and bring Coding Copilot into your next coding interview: it hears the question, reads the problem, and shows the approach and code in an overlay your screen share leaves out.
