Graviton Research Capital interview process (2026)

Round-by-round breakdown of the Graviton Research Capital software engineering interview across 3 levels and 12 rounds: interview mode, sample questions, and worked solutions.

Graviton Research Capital salary & ratings · All interview processes

Software Engineer Intern (Summer)

Difficulty: Hard · Timeline: Campus (IIT) summer-internship cycle for pre-final-year stud · Recommended prep: 3-6 months

  1. Round 0: Online Assessment (DSA) - HackerEarth

    Assessment · 5 hrs · difficulty 3/5 · elimination round

    Mode of interview: Online assessment

    Sample question: Maximum Increasing Triple Sum — Given an array A, find the maximum value of A[i] + A[j] + A[k] such that i < j < k and A[i] < A[j] < A[k]. Report -1 if no such triple exists. Input: n on the first line, then n integers. Output: the maximum triple sum, or -1.

  2. Round 1: Technical Interview 1 - DSA + Probability

    Onsite · ~45-60 min · difficulty 4/5 · elimination round

    Mode of interview: Onsite or virtual onsite interview

    Sample question: Design In-Memory File System (VS Code-style file/folder explorer) — Implement an in-memory file system from scratch (a VS Code-style file & folder explorer). Support: mkdir <path> (create a directory, creating intermediate dirs); addContentToFile <path> <content> (create file if absent, else append); readContentFromFile <path> (print file content); ls <path> (if a directory, print its immediate children in lexicographic order space-separated; if a file, print the file name). Each stdin line is one command.

  3. Round 2: Technical Interview 2 - Systems, CS Fundamentals & Puzzle

    Onsite · ~45-60 min · difficulty 3/5 · elimination round

    Mode of interview: Onsite or virtual onsite interview

    Sample question: Range Sum Query - Mutable (Segment Tree) — Support an array under two online operations: point update (set index i to value v) and range-sum query (sum of A[l..r] inclusive). Input: n, then n integers; then q; then q lines each either 'update i v' or 'sum l r'. Output one line per sum query. Intended solution is a segment tree (point update + range query) in O(log n) per op.

  4. Round 3: HR / Behavioral Round

    Behavioral · ~20-30 min · difficulty 1/5

    Mode of interview: Live behavioral interview

Software Engineer / Quantitative Developer (New Grad / SDE-1) — Gurgaon (C++, HFT)

Difficulty: Hard · Timeline: 2-4 weeks · Recommended prep: 2-3 months · Cooldown: ~6-12 month

  1. Round 0: Online Assessment (HackerEarth)

    Assessment · ~75 minutes · difficulty 3/5 · elimination round

    Mode of interview: Online assessment

    Sample question: Design a Network Packet Scheduler / Processor — Design an efficient data structure that maintains incoming network packets and processes them one by one. Packets arrive with an id and a priority; on each 'process' step, emit the highest-priority packet (ties broken by earliest arrival). Support add(id, priority) and poll() efficiently. Input: N, then N lines 'id priority' (in arrival order). Output the order in which packets are processed, space-separated.

  2. Round 1: Technical Interview 1 — Coding / CP + Resume

    Onsite · ~20-60 minutes · difficulty 4/5

    Mode of interview: Onsite or virtual onsite interview

    Sample question: Count of Smaller Numbers After Self — Given an integer array a, return an array count where count[i] is the number of elements to the right of a[i] that are strictly smaller than a[i]. The naive solution is O(n^2); the interviewer expects the divide-and-conquer (merge-sort) O(n log n) solution. Input: N then N integers. Output N integers, space-separated.

  3. Round 2: Technical Interview 2 — CS Fundamentals / Systems + Puzzles

    Onsite · ~20-60 minutes · difficulty 4/5

    Mode of interview: Onsite or virtual onsite interview

    Sample question: Range Sum Query - Mutable (Segment Tree) — Given an array, support two operations efficiently: point update (set index i to a value) and range-sum query over [l, r] inclusive. Implement with a segment tree (O(log n) per op). Input: N then N integers; then Q; then Q lines each either 'U i val' (update) or 'Q l r' (range sum). Output each query result space-separated.

  4. Round 3: Final Technical / Quant-Algorithms Round

    Onsite · ~20 minutes · difficulty 4/5

    Mode of interview: Onsite or virtual onsite interview

    Sample question: Longest Increasing Subsequence (O(n log n) tails invariant) — Given an array, return the length of the longest strictly increasing subsequence. The expected optimal is O(n log n) using the patience/tails invariant (maintain the smallest possible tail for each subsequence length). Input: N then N integers. Output a single integer.

  5. Round 4: HR / Culture-Fit & Offer

    Behavioral · ~20-30 minutes · difficulty 1/5

    Mode of interview: Live behavioral interview

Senior Software Engineer — Gurgaon (low-latency C++/HFT)

Difficulty: Hard · Timeline: 2-4 weeks · Recommended prep: Deep low-latency C++ (memory model, cach · Cooldown: No published policy (LOW).

  1. Round 0: Coding / DSA screen

    Phone · ~60 minutes · difficulty 4/5

    Mode of interview: Remote phone or video interview

    Sample question: Sliding Window Maximum — Given an array of ticks and a window size k, return the maximum of every contiguous window of size k. Must run in O(n) using a monotonic deque (no per-window rescans) — the low-latency framing is a rolling max over a stream of market-data values. Input: N then N integers then k. Output the max of each window, space-separated.

  2. Round 1: Systems / C++ architecture deep-dive

    Onsite · ~60 minutes · difficulty 4/5

    Mode of interview: Onsite or virtual onsite interview

  3. Round 2: HR / Culture-fit & Offer

    Behavioral · ~20-30 minutes · difficulty 1/5

    Mode of interview: Live behavioral interview

Other companies' interview processes

Worked sample solutions (code, design diagrams, and STAR answers) for every round are available in the FAANGPlus app: open interview processes.