Expedia ML Engineer Interview Question

Expedia ML Engineer Round 1

Q1 - Find the largest number that can be formed with the given digits

# Find the largest number that can be formed with the given digits.
# Input = [4, 7, 9, 2, 3]
# Output: 97432 

# Input = [8, 6, 0, 4, 6, 4, 2, 7]
# Output: 87664420

Q2

Expedia ML Engineer Round 2

Topics covered are ->
1. Discuss your main project in company, and what is your Technology stack?
2. Spark internals ->
1. What are the common params used while submitting Spark job?
2. How to optimize spark shuffle?
3. How does spark execute your job once you do a spark-submit?
4. Discuss challenges faced in setting up Streaming job, and what are various spark sql commands in streaming data job?
3. Spark hands-on

Assume you get input in this format in a spark dataframe, write a spark code to get What were the top 5 ranked songs in 2021? 
The output should contain columns - rank, group_name, and song_name.

year  year_rank  group_name                 artist      song_name
2021  1          Kesha                      Kesha        LMN
2021  2          Taylor                     Taylor       XYZ
2021  3          Kesha feat. Snoop Dogg     Kesha        ABC
2021  3          Kesha feat. Snoop Dogg     Snoop Dogg   ABC
2021  4          Eminem                     Eminem       DEF
2021  5          Lady Gaga feat. Rihanaa    Lady Gaga    GHI
2021  5          Lady Gaga feat. Rihanaa    Rihannaa     GHI 

Ans -
cols_list = ['year_rank', 'group_name', 'song_name']

df\
.where('year=2021 and year_rank<=5')\
.select(cols_list)\
.distinct()

Expedia ML Engineer Round 3

This is primarily System Design round.
Questions asked are ->
Discuss your main project in company, and what is your Technology stack?

Case Study
Expedia partners with Hotels to consult them to earn more revenue by pricing their facilities rightly.

Consider a use case wherein, partners want answer to below questions ->
- what is my Inventory Utilization?
Hotels allocate rooms to different vendors - Expedia, MMT, Go-Ibibo, etc.
Inventory Utilization = % of rooms utilized as proportion of number of rooms allocated to it by Hotel.
For example, if 50 rooms allocated, and 30 got booked, then inventory utilization is (30/50)*100 = 60%

- What has been my historical inventory utilization?

- How do I foresee occupancy in coming days - next 7 days, 30 days, 90 days, etc.
- This part is about applying ML Algo on predicting data.
- Here again, questions included -
- How will I store my ML Algo object?
- Where will I place my ML Algo object?
- How will I refresh it - full vs delta refresh?