*Data science interview questions:* I appeared for Google interviews couple of times , got rejected all the times but it was a good learning experience. For interview prep , I learnt a lot many new things which helped me a lot in my career. Appeared for this 1 yr back. *Designation: Data Scientist* *Coding Questions:* (they asked many questions , cant mention everything because of space constraints) 1. Given the database schema: Children(child_name STR, age INT) Toys(toy_name STR, price DEC) List(child_name STR, toy_name STR) Sample values: Children Child_Name Age -------------- Maria 5 Teo 10 Kate 8 Mark 6 Toys Toy_Name Price ---------------- PS3 200 Skate 50 Scooter 100 Lego 50 Nerf 30 List Child_Name Toy_Name ------------------- Kate PS3 Teo PS3 Teo Skate Kate Scooter Maria Lego Kate Nerf Teo Nerf Maria Skate • List (in alphabetical order) the names of the toys asked by children under 9 years old. o Follow up: How can you make the results unique/distinct? • Who are the children spending over 200? • What is the total amount spent by each child? *Case study Question :* 1. Imagine the following scenario - In 1M videos, there are 50K bad videos. There is a Model A and Model B Model A catches 75K bad videos 50K actually bad 25K good but marked bad Model B catches 25K bad videos all 25K actually bad - misses 25K bad which one would you choose and why? Tell me difference between weighted recall and F1-score ? Where to use ROC-AUC ? 2. Tell me about different frauds happening in your favorite google product (I spoke about gpay) 3. Can you tell me some methods in order to mitigate these frauds ? (Data science + Simple UI improvements / popups / awareness methods) 5. Suppose you are working as a data scientist at Meta & you see that there are users who are reporting some genuine instagram profiles as spam /fraud. It is wasting a lot of energy/time/money for inspection. How will you identify such accounts which report genuine profiles as spam with a. ML b. Non- ML approach ? Tell me strategy. *Gen AI questions :* 1. How will you solve gpay frauds with the help of Gen AI soln (take any approach and statement of your choice and ellaborate) 2. How will you improve a RAG pipeline ? what is HyDE ? 3. How will you tackle with hallucination ? 4. Can you tell me about infinte context window models ? *Statistics Questions:* 1. What is CLT ? 2. We need to draw sample from class of boys and girls staying in different zip codes , having different scores in tests. How will you draw a representative sample from this population ? 3. If your answer is stratification then how will you choose weights ? (Apart from # event \ # Total) 4. Can you tell me what is chi-square goodness of fit ? 5. What will be ditribution of gpay transactions in India ? (Need to explain : Weekdays , Weekends , Holiday seasons) *Hope it helps!* ❤️ To post your company jobs, WhatsApp at +917588467267