Plaid Software Engineer Interview Experience — A New Rule-Chain Validation Question in the Screen

Unfortunately I ran into a question that wasn't posted here before. This was a few weeks ago, and the interviewer was a woman.

Question: Chain Validation

The system has a set of validation rules. Each rule has two fields:

[rule name, result]

The result can be one of three things:
"pass" — the rule passes
"fail" — the rule fails
the name of another rule — meaning the final result of the current rule depends on that other rule

For example:

rules = [ ["rule_A", "pass"], ["rule_B", "rule_C"], ["rule_C", "fail"] ]

Here:
rule_B -> rule_C -> fail
so rule_B's final result is also fail.

You need to implement two functions.

1. isValid
   Check whether the input is valid.
   isValid(rules) -> boolean
   For each rule [name, result], result must be:
   "pass", "fail", or the name of another rule that exists in the input.
   Return true if every rule satisfies this; otherwise return false.

For example:

[ ["rule_A", "pass"], ["rule_B", "rule_C"], ["rule_C", "fail"] ]

returns:
true

While:

[ ["rule_A", "rule_X"] ]

returns false, since rule_X doesn't exist.

2. isEligible
   Check whether the entire input ultimately resolves to all pass.
   isEligible(rules) -> boolean
   If a rule's result is another rule, you need to keep following the reference until you get either:
   pass or fail.
   Only return true if every rule ultimately resolves to pass.

For example:

[ ["rule_A", "pass"], ["rule_B", "rule_C"], ["rule_C", "pass"] ]

Resolves to:
rule_A -> pass
rule_B -> rule_C -> pass
rule_C -> pass
So: isValid -> true, isEligible -> true

If instead:

[ ["rule_A", "pass"], ["rule_B", "rule_C"], ["rule_C", "fail"] ]

then:
rule_B -> rule_C -> fail
so: isValid -> true, isEligible -> false

Addendum (2026-08-20 01:21 +08:00): I interviewed for the mid-level role. The first two rounds were coding, and another round was a question that's already been posted here before.