Arcesium Lead Engineer Interview Experience (Couldn't Clear Round 2)
- Posted
- 3 Jun 2026
Summary
I interviewed for a Lead Engineer position at Arcesium, completed an online assessment and a DSA round, but did not clear Round 2.
How it went
Round 1 - Online Assessment
Question 1: Expression Evaluator
Given JSON input from an API endpoint containing expression groups, evaluate expressions based on their type:
- DIRECT
- DOLLAR_EXPRESSION
- RS_EXPRESSION
Requirements included:
- Dependency resolution between expressions
- Variable substitution using
${variable} - Extensible design for future expression types
- Memory-constrained processing (millions of groups, limited groups in memory at a time)
The focus seemed to be on clean design, dependency handling, and scalability.
Round 2 - DSA
Question 1: Decode Numeric String to Letters
Encoding:
A → 1 B → 2 ... Z → 26
Example:
2116
Possible decodings:
- 2,1,1,6
- 21,1,6
- 2,11,6
- 2,1,16
- 21,16
Answer = 5
Essentially a variation of the classic Decode Ways problem.
Question 2: Check Ancestor Relationship in Tree
Given a tree and multiple queries (X, Y), determine whether X is an ancestor of Y.
Example:
1 -> 2 1 -> 3 2 -> 4 2 -> 5 3 -> 6 5 -> 7
Queries:
1 6 => true 6 1 => false 2 7 => true 2 6 => false
Takeaways
- DSA fundamentals are still extremely important even for senior/lead roles.
- Interviewers were interested in both correctness and discussion around complexity.
- Tree traversal and DP-based problems remain common.
I didn't make it through this time, but it was a good learning experience. Hope this helps someone preparing for Arcesium interviews.
Good luck to everyone interviewing!
Questions asked (3)
- 01
Expression Evaluator
System Design
Given JSON input from an API endpoint containing expression groups, evaluate expressions based on their type (DIRECT, DOLLAR_EXPRESSION, RS_EXPRESSION). Requirements include dependency resolution between expressions, variable substitution using
${variable}, extensible design for future expression types, and memory‑constrained processing (handling millions of groups with limited groups in memory at a time). The focus is on clean design, dependency handling, and scalability. - 02
Decode Numeric String to Letters
Data Structures & Algorithms · Medium
Encoding: A → 1, B → 2, …, Z → 26. Given a numeric string (e.g., "2116"), count all possible decodings into letters. Example decodings for "2116": 2,1,1,6; 21,1,6; 2,11,6; 2,1,16; 21,16. The answer for the example is 5. This is a variation of the classic Decode Ways problem.
- 03
Check Ancestor Relationship in Tree
Data Structures & Algorithms · Medium
Given a tree and multiple queries (X, Y), determine whether X is an ancestor of Y. Example tree edges: 1->2, 1->3, 2->4, 2->5, 3->6, 5->7. Sample queries: 1 6 => true, 6 1 => false, 2 7 => true, 2 6 => false.