Prepare for and practice interview questions from Tsmc.

Tsmc Interview Questions

Tsmc Interview Guides

Machine Learning

How do we deal with the missing square footage data to construct our model

Missing Housing Data

Why can the same machine learning algorithm produce different success rates on the same dataset

Same Algorithm Different Success

What are some reasons why measuring bias would be important in building this particular model

Food Delivery Times

How would you evaluate whether a decision tree is the right model and assess its performance before and after deployment

Decision Tree Evaluation

How does random forest generate the forest and why use it over logistic regression

Random Forest Explanation

Data Structures & Algorithms

Find the missing integer from a array of consequtive integers

Find the Missing Number

Write a function one_element_removed that returns the integer removed from list Y in O(1) space and O(n) time without using set

One Element Removed

Write a function to return True if two strings are anagrams of each other and False if they are not

Valid Anagram

Implement a fixed-length array with addition, deletion, and search operations.

Fixed-Length Arrays: Deletion

Write a function that returns a boolean indicating if a value is in the linked list.

Search Linked List

Statistics

What are the assumptions of linear regression?

Assumptions of Linear Regression

What is the downside of only using the R-Squared (R^2) value to determine a relationship between two variables

Using R Squared

Explain the difference between covariance and correlation and provide an example

Covariance vs Correlation

How do you detect and handle correlation between variables in linear regression, and what happens if you ignore it

Correlation in Regression

What does it mean to "bootstrap" a data set?

Bootstrapping Samples

Behavioral

Describing a data project and its challenges

Hurdles In Data Projects

Strategically resolving misaligned expectations with stakeholders for a successful project outcome

Stakeholder Communication

What do you tell an interviewer when they ask you what your strengths and weaknesses are?

Your Strengths and Weaknesses

Analytics

Create and write queries for health metrics for stack overflow 

Community Health Metrics

Challenges of specific student test score layouts, recommended formatting changes for enhanced analysis, and common issues found in "messy" datasets.

Digitizing Student Test Scores

Describing a real-world data cleaning and organization project

Data Cleaning Experiences

You have an array of integers, <code>nums</code> of length <code>n</code> spanning <code>0</code> to <code>n</code> with one missing. Write a function <code>missing_number</code> that returns the missing number in the array.

Note: Complexity of $O(n)$ required.

Example:

Input:

<pre tabindex="0" class="chroma"><code>nums = [0,1,2,4,5] 
missing_number(nums) -&gt; 3
</code></pre>

There are two lists, <code>list X</code> and <code>list Y</code>. Both lists contain integers from <code>-1000</code> to <code>1000</code> and are identical to each other except that one integer is removed in <code>list Y</code> that exists in <code>list X</code>.

Write a function <code>one_element_removed</code> that takes in both lists and returns the integer that was removed in $O(1)$ space and $O(n)$ time without using the python set function.

Example:

Input:

<pre tabindex="0" class="chroma"><code>list_x = [1,2,3,4,5]
list_y = [1,2,4,5]

one_element_removed(list_x, list_y) -&gt; 3
</code></pre>

Which of the following methods would be the most efficient way of identifying what element is not included in two lists of integers that differ by a single element being removed?

For example, if the lists are <code>[1,2,3,4,5]</code> and <code>[1,2,4,5]</code>, the function should return <code>3</code>

Which of the following methods would be the most efficient way of identifying what element is not included in two lists of integers that differ by a single element being removed? 

For example, if the lists are `[1,2,3,4,5]` and `[1,2,4,5]`, the function should return `3`

Efficiently Identifying Missing Integer

Say you are tasked with analyzing how well a model fits the data given. You want to determine a relationship between two variables.

What is the downside of only using the R-Squared $(R^2)$ value to do so?

In the context of regression analysis, why is it misleading to rely solely on a high R-squared ($R^2$) value to infer a causal relationship between variables?

In the context of regression analysis, why is it misleading to rely solely on a high R-squared ($R^2$) value to infer a causal relationship between variables?

Using R Squared I

Which of the following best describes the R-squared ($R^2$) value in the context of a regression model?

Which of the following best describes the R-squared ($R^2$) value in the context of a regression model?

Using R Squared II

What is a limitation of using R-squared ($R^2$) as the only metric to evaluate a regression model's fit?

What is a limitation of using R-squared ($R^2$) as the only metric to evaluate a regression model's fit?

Using R Squared III

What is the difference between covariance and correlation? Provide an example.

Given a <code>ListNode</code> representing the head of a linked list, write a function <code>is_cyclic</code> which returns <code>True</code> if a linked list is cyclic and <code>False</code> if it is not. A linked list is cyclic when there is no tail to the linked list, and the supposed tail is attached to another node inside the list itself, creating a cycle.

A <code>ListNode</code> is defined as:

<pre tabindex="0" class="chroma"><code>class ListNode:
	value: ListNode = None
 next: ListNode = None
</code></pre>

Example:

Input:

<img src="/s3/questions/1a65651d-a483-4481-b073-2738d65ab9d7.png" alt="image"/>

Output:

<pre tabindex="0" class="chroma"><code>def is_cyclic(head) -&gt; True
</code></pre>

Cyclic Detection

Why would the same machine learning algorithm generate different success rates using the same dataset?

Note: When they ask us an ambiguous question, we need to gather context and restate it in a way that’s clear for us to answer.

We want to build a model to predict housing prices in the city of Seattle. We’ve scraped 100K sold listings over the past three years but found that around 20% of the listings are missing square footage data.

How do we deal with the missing data to construct our model?

We want to build a model to predict housing prices in the city of Seattle. We’ve scraped 100K sold listings over the past three years, but have discovered that around 20% of the listings are missing square footage data.

How would you approach dealing with this missing data, in order to construct the most useful predictive model possible?

We want to build a model to predict housing prices in the city of Seattle. We’ve scraped 100K sold listings over the past three years, but have discovered that around 20% of the listings are missing square footage data.

How would you approach dealing with this missing data, in order to construct the most useful predictive model possible?



Three zebras are chilling in the desert. Suddenly a lion attacks.
Each zebra is sitting on a corner of a triangle with sides of equal length.
Each zebra randomly picks a direction and only runs along the outline of the triangle to either edge of the triangle.

What is the probability that none of the zebras collide?

Three zebras are chilling in the desert. Suddenly a lion attacks.

Each zebra is sitting on a corner of a triangle with side of equal length. Each zebra randomly picks a direction and only runs along the outline of the triangle to either edge of the triangle.

What is the probability that none of the zebras collide?

Three zebras are chilling in the desert. Suddenly a lion attacks.

Each zebra is sitting on a corner of a triangle with side of equal length.  Each zebra randomly picks a direction and only runs along the outline of the triangle to either edge of the triangle.

What is the probability that none of the zebras collide?



Three Zebras

Brainteasers

An electronics retail company tracks its sales and wants to run a performance analysis. They’re hoping to categorize sales into different groups, based on the amount of sales and the region where the sale occurred. The conditions for categorization are:

<ul>
<li>Standard Sales: Any sales below 2000 are considered standard, except if they occur in the <code>East</code> region, or in the month of July.</li>
<li>Premium Sales: Any sales that are over or equal to 2000, or any sales in the <code>East</code> region (regardless of the amount), are considered premium, except again during the month of July.</li>
<li>Promotional Sales: All sales made in the promotional period of July, regardless of the amount or region, are considered promotional.</li>
</ul>

The company wants a report that sums these amounts by region to help with financial planning and bonus allocations.

Note: You may assume that table contains data only for 2023

Example:

Input:

<code>sales</code> table

<table>
<thead>
<tr>
<th>Column</th>
<th>Type</th>
</tr>
</thead>

<tbody>
<tr>
<td><code>sale_id</code></td>
<td>INTEGER</td>
</tr>

<tr>
<td><code>employee_id</code></td>
<td>INTEGER</td>
</tr>

<tr>
<td><code>product_id</code></td>
<td>INTEGER</td>
</tr>

<tr>
<td><code>sale_amount</code></td>
<td>FLOAT</td>
</tr>

<tr>
<td><code>sale_date</code></td>
<td>DATE</td>
</tr>

<tr>
<td><code>region</code></td>
<td>TEXT</td>
</tr>
</tbody>
</table>
Output:

<table>
<thead>
<tr>
<th>Column</th>
<th>Type</th>
</tr>
</thead>

<tbody>
<tr>
<td><code>region</code></td>
<td>TEXT</td>
</tr>

<tr>
<td><code>total_sales</code></td>
<td>FLOAT</td>
</tr>

<tr>
<td><code>premium_sales</code></td>
<td>FLOAT</td>
</tr>

<tr>
<td><code>standard_sales</code></td>
<td>FLOAT</td>
</tr>

<tr>
<td><code>promotional_sales</code></td>
<td>FLOAT</td>
</tr>
</tbody>
</table>

Categorize sales based on the amount of sales and the region

Categorize Sales

Given a 2-D matrix P of predicted values and actual values, write a function precision_recall to calculate precision and recall metrics.

Return the ordered pair (precision, recall).

Example:

Input:

<pre tabindex="0" class="chroma"><code>
P = [[121, 9],
 [17, 144]]
</code></pre>

where P represents the following table

<table>
<thead>
<tr>
<th></th>
<th>Predicted =TRUE</th>
<th>Predicted =FALSE</th>
</tr>
</thead>

<tbody>
<tr>
<td>Actual =TRUE</td>
<td>121</td>
<td>9</td>
</tr>

<tr>
<td>Actual =FALSE</td>
<td>17</td>
<td>144</td>
</tr>
</tbody>
</table>
Output:

<pre tabindex="0" class="chroma"><code>def precision_recall(P) -&gt; (0.931, 0.877,)
</code></pre>

Write a function to calculate precision and recall metrics.

Precision and Recall

Tell me about a data project that didn’t go the way you expected. What did you set out to do, what surprised you, and how did you handle it?

What are the assumptions of linear regression?

Which assumption of the residuals of the standard linear regression model can not be overcome by increasing the sample size?

Regression assumptions

You’re a data scientist at YouTube. Leadership wants to revisit how the platform measures success, not for a specific new feature, but for the product as a whole given its current state.

How would you define and measure success for YouTube? Identify your primary and secondary metrics, explain your rationale for each, and discuss how you would choose the right thresholds for any engagement-based metrics you propose.

The role of A/B testing in measuring the success rate of an analytics experiment

Success Measurement

Product Sense & Metrics

Given two strings, write a function to return <code>True</code> if the strings are anagrams of each other and <code>False</code> if they are not.

Note: A word is not an anagram of itself.

Example 1:

Input:

<pre tabindex="0" class="chroma"><code>string_1 = &#34;listen&#34;
string_2 = &#34;silent&#34;
</code></pre>

Output:

<pre tabindex="0" class="chroma"><code>True
</code></pre>

Example 2:

Input:

<pre tabindex="0" class="chroma"><code>string_1 = &#34;banana&#34;
string_2 = &#34;bandana&#34;
</code></pre>

Output:

<pre tabindex="0" class="chroma"><code>False
</code></pre>

Write a function, <code>search_list</code>, that returns a boolean indicating if the <code>target</code> value is in the <code>linked_list</code> or not.

You receive the head of the linked list, which is a dictionary with the following keys:
- <code>value</code>: contains the value of the node.
- <code>next</code>: contains the next node in the list, or <code>None</code>.

If the linked list is empty, you’ll receive <code>None</code> since there is no head node for an empty list.

Example:

Input:

<pre tabindex="0" class="chroma"><code>target = 2
linked_list = 3 -&gt; 2 -&gt; 5 -&gt; 6 -&gt; 8 -&gt; None
</code></pre>

Output:

<pre tabindex="0" class="chroma"><code>search_list(target, linked_list) -&gt; True
</code></pre>

Assume that you have data on student test scores, and they’re in one of the following layouts (dataset 1 and dataset 2).

What are the drawbacks of having the data organized in such a way? What formatting changes would you make so the data is more useful for analysis?

Additionally, describe some of the common problems seen in “messy” datasets.

Dataset 1 and 2:

<img src="/s3/questions/a69a2739-0c3e-4841-b1a0-83f8c07c35e1.png" alt="Messy Dataset"/>

How does random forest generate the forest? Additionally, why would we use it over other algorithms such as logistic regression?

What happens when you average the output on multiple decision trees?

What happens when you average the output on multiple decision trees?

Average Trees

Suppose an array sorted in ascending order is rotated at some pivot unknown to you beforehand.

You are given a target value to search. If the value is in the array, then return its index; otherwise, return -1. 

Notes:

<ul>
<li>Rotating an array at pivot $n$ gives you a new array that begins with the elements after position $n$ and ends with the elements up to position $n$.</li>
<li>You may assume no duplicate exists in the array.</li>
</ul>

Bonus: Your algorithm’s runtime complexity should be in the order of $O(\log n)$.

Example:

Input: 

<pre tabindex="0" class="chroma"><code>sample_input = [0,1,2,4,5,6,7] 
rotated_input = [4,5,6,7,0,1,2]
target_value = 6
</code></pre>

Output:

<pre tabindex="0" class="chroma"><code>def target_value_search(rotated_input, target_value) -&gt; 2
</code></pre>

Search for a value in log(n) over a sorted array that has been shifted.

Target Value Search

Imagine you are asked to build a machine learning model to decide new loan approvals for a
financial firm. You ask the data department in the company for a subset of data to get started
working on the problem. The data includes different features about applicants such as age,
occupation, zip code, height, number of children, favorite color, etc. You decide to build
multiple machine learning models to test out different ideas before settling on the best one.

How would you explain the bias-variance tradeoff with regards to
building and choosing a model to use?

Bias vs. Variance Tradeoff

How would you handle the data preparation for building a machine learning model using imbalanced data?

Addressing imbalanced data in machine learning through carefully prepared techniques.

Data Preparation for Imbalanced Data

Let’s say you work as a machine learning engineer at Intel, focusing on AI-powered document analysis for enterprise clients. You are tasked with building a system that retrieves and summarizes relevant information from technical documents, some of which exceed the context window of your deployed language model.

How would you design an end-to-end retrieval system to efficiently extract and summarize information from these oversized documents?

Design a retrieval system for extracting and summarizing information from documents larger than a model's context window at Intel.

Oversized Document Retrieval

AI & Agentic Systems

Let’s say that you’re training a classification model.

How would you combat overfitting when building tree-based models?

Let's say that you're training a classification model.   How would you combat overfitting when building tree-based models?

Overfit Avoidance

Create an <code>Array</code> class simulating the functionality of fixed-size arrays. The array’s size is <code>6</code>. The array’s constructor is already defined:

<pre tabindex="0" class="chroma"><code>class Array:

 def __init__(self):
 pass # Dynamically added to avoid tampering
 self._count = 0
 self._array = [None, None, None, None, None, None]
 self._MAX_CAPACITY = 6
</code></pre>

There is no need to implement the constructor. In the code editor, <code>__init__</code> will contain <code>pass</code>. The constructor’s code will be added dynamically to avoid any changes.

The <code>Array</code> class should have the following methods:

<ul>
<li><code>__len__()</code>: Returns the length of the array. For example, if the array is <code>[1, 2, None, None, None, None, None]</code>, <code>__len__()</code> will return <code>2</code>.</li>

<li><code>__getitem__(index)</code>: Returns the element at the specified index. Raises an <code>IndexError</code> when the index is out of range.</li>

<li><code>emplace_back(element)</code>: Places an element at the back of the array. Raises an <code>ArrayFull</code> exception when the array is full (length &gt; 6).</li>

<li><code>emplace_front(element)</code>: Places an element at the front of the array. Raises an <code>ArrayFull</code> exception when the array is full (length &gt; 6).</li>

<li><code>emplace(element, index)</code>: Places an element at the specified index. Raises an <code>ArrayFull</code> exception when the array is full (length &gt; 6).</li>
</ul>

Note: <code>ArrayFull</code> is already implemented. To raise it, simply use <code>raise ArrayFull()</code>.

Implement the addition operations of fixed length arrays.

Fixed Length Arrays: Addition

Given a string, write a function to determine if it is palindrome or not.

Note: A palindrome is a word/string that is read the same way forward as it is backward, e.g. <code>&#39;reviver&#39;</code>, <code>&#39;madam&#39;</code>, <code>&#39;deified&#39;</code> and <code>&#39;civic&#39;</code> are all palindromes, while <code>&#39;tree&#39;</code>, <code>&#39;music&#39;</code> and <code>&#39;person&#39;</code> are not palindromes.

Example:

Input:

<pre tabindex="0" class="chroma"><code>word1 = &#34;tree&#34;
word2 = &#34;radar&#34;
</code></pre>

Output:

<pre tabindex="0" class="chroma"><code>def is_palindrome(word1) -&gt; False
def is_palindrome(word2) -&gt; True
</code></pre>

Given a string, write a function to determine if it is palindrome or not.

String Palindromes

What are time series models? Why do we need them when we have less complicated regression models?

What can happen if a traditional regression model is used on data that has autocorrelation?

What can happen if a traditional regression model is used on data that has autocorrelation?

Why Do We Need Time Series Models? I

Why can't standard regression models be used effectively for data points where current values are influenced by past values?

Why can't standard regression models be used effectively for data points where current values are influenced by past values?

Why Do We Need Time Series Models? II

Explain what time series models are and why they’re needed instead of simpler regression models

Why Do We Need Time Series Models?

Forecasting & Time Series

When an interviewer asks a question along the lines of:

<ul>
<li>What would your current manager say about you? What constructive criticisms might he give?</li>
<li>What are your three biggest strengths and weaknesses you have identified in yourself?</li>
</ul>

How would you respond?

When asked about your strengths in an interview, what is an effective way to respond?

When asked about your strengths in an interview, what is an effective way to respond?

Your Strengths and Weaknesses I

Which of the following is an acceptable strategy when discussing weaknesses in an interview?

Which of the following is an acceptable strategy when discussing weaknesses in an interview?

Your Strengths and Weaknesses II

You work at Stack Overflow on the community team that monitors the health of the platform.

Community members can create a post to ask a question, and other users can reply with answers or comments to that question. The community can express their support for the post by upvoting or downvoting.

Example:

Input:

<code>post_analytics</code> table

<table>
<thead>
<tr>
<th>Column</th>
<th align="center">Type</th>
<th align="center">Description</th>
</tr>
</thead>

<tbody>
<tr>
<td><code>id</code></td>
<td align="center">INTEGER</td>
<td align="center">Primary key of posts table</td>
</tr>

<tr>
<td><code>user_id</code></td>
<td align="center">INTEGER</td>
<td align="center">ID of the user who created the post</td>
</tr>

<tr>
<td><code>created_at</code></td>
<td align="center">DATETIME</td>
<td align="center">Timestamp of the post</td>
</tr>

<tr>
<td><code>title</code></td>
<td align="center">VARCHAR</td>
<td align="center">Title of the post</td>
</tr>

<tr>
<td><code>body</code></td>
<td align="center">VARCHAR</td>
<td align="center">Text content of the post</td>
</tr>

<tr>
<td><code>comment_count</code></td>
<td align="center">INTEGER</td>
<td align="center">Total number of the comments on a post</td>
</tr>

<tr>
<td><code>view_count</code></td>
<td align="center">INTEGER</td>
<td align="center">Total number of the views on a post</td>
</tr>

<tr>
<td><code>answer_count</code></td>
<td align="center">INTEGER</td>
<td align="center">Total number of answers on a post</td>
</tr>

<tr>
<td><code>upvotes</code></td>
<td align="center">INTEGER</td>
<td align="center">Total number of upvotes on the post</td>
</tr>
</tbody>
</table>

<ol>
<li>Given a table of Stack Overflow posts data, suggest three metrics to monitor the health of the community.</li>

<li>Write the queries for these metrics in SQL.</li>
</ol>

Let’s say that you work as a data scientist at a food delivery company.

You are tasked with building a model to predict food preparation times at a restaurant from the time when a customer sends in the order to when the meal is ready for the driver to pick up.

What are some reasons why measuring bias would be important in building this particular model?

Write a function <code>fib</code> which takes an integer <code>n</code> and returns the nth fibonacci number. For this question, <code>fib</code> MUST be recursively defined.

Note: <code>for</code> and <code>while</code> keywords are disabled for this question, as well as <code>functools</code>. Any workarounds around this restriction is prohibited.

Example:

Input:

<pre tabindex="0" class="chroma"><code>n = 4
</code></pre>

Output:

<pre tabindex="0" class="chroma"><code>def fib(n) -&gt; 3
# the sequence is [1, 1, 2, 3, 5, 8...]
</code></pre>

Implement a dynamic recursive fibonacci function.

Impossibly Iterative Fibonacci

Find if there is a path from a starting point to an ending point in a walled maze

Calculate Moving Average	SQL	Easy
Predict Customer Churn	Machine Learning	Medium
A/B Test Significance	Statistics	Medium
Optimize Query Performance	SQL	Hard
Feature Importance Analysis	Machine Learning	Medium
Clean Missing Data	Python	Easy
Neural Network Architecture	Deep Learning	Hard
Calculate Cohort Retention	SQL	Medium
Bayesian Probability	Statistics	Easy
Recommend Similar Products	Machine Learning	Hard

Tsmc Interview Questions

Tsmc Interview Guides

Tsmc Interview Questions

Challenge

Tsmc Salaries by Position

Discussion & Interview Experiences

Discussion & Interview Experiences