Autocoder

Arcesium Interview Questions and Answers

Last updated 2026-11-19

Arcesium builds data and operations technology for investment managers, so its engineering interviews commonly cover data structures and algorithms alongside databases, object-oriented design and the reliability of systems that handle financial data. This page offers practice questions in that style with worked answers.

They are practice questions on the topics the process is commonly reported to cover, not real papers. For timed problems see Arcesium coding round questions, and for the SQL topics see SQL interview questions.

Rounds as commonly reported

Candidate accounts describe an online assessment with aptitude and coding problems, then two or three technical interviews mixing DSA, DBMS and SQL, OOP and a project discussion, and a final managerial or HR conversation. Some reports also mention puzzles. Arcesium has not published a fixed process, so treat this as a pattern and rely on your recruiter's information.

Topics worth revising

Trees, graphs, dynamic programming and heaps; SQL joins, aggregates, window functions; transactions, indexing and isolation levels; OOP design and the SOLID principles; and a brief, clear explanation of your projects.

Hiring status

We have no sourced hiring dates, numbers or process facts for Arcesium and no /hiring page for it on this site, so we state none. Check the company's own careers site for current roles and the application route.

Arcesium Interview Questions and Answers: questions with worked answers

  1. 1

    Explain the ACID properties and the isolation levels.

    medium

    Atomicity: a transaction happens fully or not at all. Consistency: it moves the database from one valid state to another. Isolation: concurrent transactions do not see each other's partial work. Durability: committed data survives crashes. Isolation levels trade anomalies for concurrency: read uncommitted allows dirty reads; read committed prevents them; repeatable read also prevents non-repeatable reads; serializable prevents phantoms as well. For financial data, interviewers like to hear why you would accept the cost of a stronger level for balances.

  2. 2

    Write SQL to find the second highest salary.

    easy

    Use DENSE_RANK over salary descending and filter rank 2, which handles ties correctly. Alternatively, select the maximum salary below the overall maximum. Return NULL when there is no second value; the subquery form does this naturally.

    sql
    SELECT MAX(salary) AS second_highest
    FROM employees
    WHERE salary < (SELECT MAX(salary) FROM employees);
  3. 3

    Find duplicate email addresses in a users table.

    easy

    Group by the column and keep groups with more than one row. To delete duplicates while keeping the lowest id, join the table to itself or use a window function to number rows per email and delete rows numbered above one.

    sql
    SELECT email, COUNT(*) AS n
    FROM users
    GROUP BY email
    HAVING COUNT(*) > 1;
  4. 4

    How does a B+ tree index speed up queries, and when does it not help?

    medium

    A B+ tree keeps keys sorted in balanced nodes with a high fan-out, so a lookup reads only a few pages (logarithmic in the number of rows), and leaves are linked for fast range scans. It does not help when the predicate is not sargable (a function on the column, a leading wildcard in LIKE), when a query returns most of the table, or on very low-cardinality columns. Every index also slows writes and takes space.

  5. 5

    Design a position and trade tracker for a portfolio.

    hard

    Entities: Instrument, Portfolio, Trade (immutable, with id, instrument, quantity, price and timestamp) and Position (derived from trades). Store trades append-only and compute positions by aggregation or maintain them as a projection updated on each trade; this gives an audit trail. Handle corrections with reversing trades, never updates. Use decimal types, not floating point, for money. For scale, partition trades by portfolio and use idempotency keys so a retried message does not double-count.

  6. 6

    Explain the SOLID principles briefly with an example.

    medium

    Single responsibility: separate report generation from report delivery. Open/closed: add a new fee calculator by adding a class, not editing a switch statement. Liskov substitution: a subclass must honour the base class contract. Interface segregation: many small interfaces beat one large one. Dependency inversion: depend on a Notifier interface, not a concrete email sender, which makes tests simple.

  7. 7

    Eight balls, one heavier, a balance scale. Find it in two weighings.

    easy

    Weigh three against three. If they balance, the heavy ball is among the remaining two: weigh those against each other. If one side is heavier, take that group of three, weigh two of them: if one is heavier it is the one, if they balance it is the third. Two weighings suffice.

  8. 8

    Detect a cycle in a directed graph.

    medium

    Use DFS with three colours: white (unvisited), grey (on the current path) and black (done). An edge to a grey node means a cycle. Alternatively run Kahn's topological sort and check whether all nodes were removed. O(V + E).

    python
    def has_cycle(n, edges):
        adj = [[] for _ in range(n)]
        for u, v in edges:
            adj[u].append(v)
        color = [0] * n
        def dfs(u):
            color[u] = 1
            for v in adj[u]:
                if color[v] == 1 or (color[v] == 0 and dfs(v)):
                    return True
            color[u] = 2
            return False
        return any(color[i] == 0 and dfs(i) for i in range(n))
  9. 9

    Difference between HashMap and ConcurrentHashMap in Java.

    medium

    HashMap is not thread safe: concurrent writes can lose updates or corrupt state. ConcurrentHashMap allows many threads at once, using CAS and per-bin locking in modern versions, so reads rarely block, and it provides atomic operations such as putIfAbsent, compute and merge. It disallows null keys and values. Collections.synchronizedMap locks the whole map for every call, which scales poorly.

  10. 10

    What is normalisation, and when would you denormalise?

    medium

    Normalisation removes redundancy by splitting data so each fact lives in one place (1NF atomic values, 2NF no partial dependency on a composite key, 3NF no transitive dependency), which avoids update anomalies. Denormalise deliberately for read-heavy reporting where joins are too costly, accepting duplicated data that must be kept consistent, often through pipelines or materialised views.

  11. 11

    Tell me about a time you found a bug in production or in someone else's work.

    easy

    Describe how you detected it (a metric, a customer complaint, a test), how you contained the impact first, then how you fixed the root cause and added a test or alert. Emphasise communication: who you told and when. Where the bug was yours, say so plainly. Reliability matters a great deal when software touches financial records.

  12. 12

    Why Arcesium, and why this role?

    easy

    Connect the business to the engineering: data-heavy systems for investment managers, where correctness and reliability matter. Show you have read about what the company does and can name a technical area you want to learn, such as large-scale data pipelines or financial domain modelling.

Frequently asked questions

Does Arcesium ask SQL?

SQL and database fundamentals are commonly reported for engineering roles. Revise joins, grouping and window functions.

Are there puzzles in the interview?

Some candidate accounts mention puzzles. Prepare a few standard ones and, more importantly, practise explaining your reasoning.

What coding difficulty is typical?

Commonly reported as medium, with some harder problems in later rounds. See Arcesium coding round questions.

Where can I check current openings?

On Arcesium's own careers site. This site has no sourced hiring data for the company.

Related guides

More in Company Interview Questions

See all in this section