InterviewDB Experience · New York

Pronunciation Annotation - Tag Words in Text with Phonetic Metadata

phone
Sign in to bookmark

Interview Experience

Round 1 Coding

Problem

Given a sentence and a dictionary mapping words to their phonetic annotation (e.g. IPA or a simplified pronunciation string),

return the sentence with each recognized word wrapped in an annotation tag. Unrecognized words are left as-is. The matching is case-insensitive; the original casing must be preserved in output.

python
def annotate_pronunciation(
    sentence: str,
    phoneme_dict: dict[str, str]
) -> str:
    pass

Example


**Input**:
sentence = "The quick brown fox"
phoneme_dict = {"the": "dh-ah", "fox": "f-aa-k-s"}

**Output**:
"[The|dh-ah] quick brown [fox|f-aa-k-s]"

Example 2


**Input**:
sentence = "Reading is fun"
phoneme_dict = {"reading": "r-ee-d-ih-ng"}

**Output**:
"[Reading|r-ee-d-ih-ng] is fun"

Follow-ups

  1. How do you handle multi-word phrases in the dictionary, like {"New York": "n-y-oo-y-aw-r-k"}?
  2. What if the same word has different pronunciations depending on part of speech (e.g. "read" present vs past)?
  3. How would you build the phoneme_dict automatically from a corpus using a text-to-speech API?
  4. How do you handle punctuation attached to words (fox, should still match fox)?

Full Details

Round 1 Coding

Problem

Given a sentence and a dictionary mapping words to their phonetic annotation (e.g. IPA or a simplified pronunciation string),

return the sentence with each recognized word wrapped in an annotation tag. Unrecognized words are left as-is. The matching is case-insensitive; the original casing must be preserved in output.

python
def annotate_pronunciation(
    sentence: str,
    phoneme_dict: dict[str, str]
) -> str:
    pass

Example


**Input**:
sentence = "The quick brown fox"
phoneme_dict = {"the": "dh-ah", "fox": "f-aa-k-s"}

**Output**:
"[The|dh-ah] quick brown [fox|f-aa-k-s]"

Example 2


**Input**:
sentence = "Reading is fun"
phoneme_dict = {"reading": "r-ee-d-ih-ng"}

**Output**:
"[Reading|r-ee-d-ih-ng] is fun"

Follow-ups

  1. How do you handle multi-word phrases in the dictionary, like {"New York": "n-y-oo-y-aw-r-k"}?
  2. What if the same word has different pronunciations depending on part of speech (e.g. "read" present vs past)?
  3. How would you build the phoneme_dict automatically from a corpus using a text-to-speech API?
  4. How do you handle punctuation attached to words (fox, should still match fox)?
Free preview. Unlock all Heygen questions →

About This Question

This is a candidate experience report from a heygen interview during the phone round.

It covers the following topics: Coding, Phone, Onsite, Strings .

Topics

Coding Phone Onsite Strings
← All Heygen questions Heygen Software Engineer Questions → All Strings Questions →