Week 3 - Prompt Engineering & Productivity : Copilot Free learning & cert prep

[Akash1134]

👋 Welcome to Week 3 of the GitHub Copilot Free learning journey and cert prep! We're now at the halfway point in our learning journey, and it’s been incredible to see the depth of thought and creativity in your submissions so far.

This week, we’re shifting our focus to some of the core principles of prompt engineering and productivity. Understanding these foundational concepts will help you craft more effective prompts, allowing you to harness Copilot’s full potential for smarter, faster coding. Here's what we'll be covering to refine your skills and take your workflow to the next level 🚀

• Prompt Crating and Prompt Engineering
  • Fundamentals of prompt crafting
  • Fundamentals of prompt engineering
• Developer use cases for AI
  • Improve developer productivity
  • How GitHub Copilot can help with SDLC
  • Limitations of using GitHub Copilot
  • How to use the productivity API to see how GitHub Copilot impacts coding

Note

Don’t miss out—Keep engaging, learning, and growing together! Top participants will earn a GitHub Certifications exam voucher! 🎫 Think of it as your golden ticket to advancing your career! 💼

Your First Move: Study Smart, Level Up! 🎯

We've already explored Copilot’s features and data handling, giving you a solid foundation in how it works. Now, it's time to take your skills to the next level by mastering the art of prompt creation and prompt engineering. The way you frame your inputs directly impacts the quality of AI-generated outputs, and learning how to craft effective prompts will help you get the most out of Copilot. We'll also dive into real-world developer use cases, showing you practical ways Copilot can streamline your workflow, boost productivity, and enhance your coding experience.

• Microsoft Learn Modules (each of the modules has its own knowledge check exercise)
  • Introduction to prompt engineering with GitHub Copilot
  • Developer use cases for AI with GitHub Copilot
    • Using GitHub Copilot with JavaScript (Optional)
    • Using GitHub Copilot with Python (Optional)
  • Implement code improvements using GitHub Copilot tools
  • Develop code features using GitHub Copilot tools
• GitHub Learning Pathways
  • Essentials of GitHub Copilot (series)
• 🤩 Microsoft Ignite 2024 (Bonus)
  • Maximize developer efficiency with GitHub Copilot for partner success
  • Top 5 GitHub Copilot Features at Microsoft Ignite 2024
  • Discover GitHub Copilot's untapped potential in Visual Studio
  • GitHub Copilot's Game Changing Features
• Videos
  • Master the core principles of prompt engineering with GitHub Copilot
  • Visual Studio Prompt Engineering with GitHub Copilot
  • GitHub Copilot Study Guide Series 2: Developer use cases for AI with GitHub Copilot
  • Level up with GitHub Copilot: using AI to learn, code, and build
  • Getting started with GitHub Copilot (Playlist with over 80 videos)
• Blogs and Resources
  • Research: quantifying GitHub Copilot’s impact on developer productivity and happiness
  • What’s a prompt and what is prompt engineering?
  • What are the pros and cons of using GitHub Copilot?
  • How to use GitHub Copilot: Prompts, tips, and use cases
• GitHub Docs
  • Prompt engineering for GitHub Copilot
  • Use GitHub Copilot
  • Manage GitHub Copilot

Knowledge Checkpoint 🏁 - Let’s See What You’ve Got! 🧠

1. When using GitHub Copilot in a corporate environment, which of the following practices could introduce legal or security risks? (Select two)

A) Using Copilot to generate code for encryption algorithms without verification.
B) Enabling Copilot to suggest code completions within a private enterprise repository.
C) Accepting Copilot-generated code that lacks licensing information.
D) Using Copilot in an IDE that does not support secure API calls.

2. A developer is working with Copilot in Visual Studio and wants to optimize Copilot’s inline chat for debugging. What is the best way to achieve this?

A) Rewriting prompts to be more detailed, explicitly mentioning expected outputs.
B) Running Copilot Chat multiple times to get varied debugging suggestions.
C) Adjusting Copilot’s temperature parameter to control randomness in responses.
D) Restricting Copilot Chat to only use file-specific context rather than project-wide scope.

3. A developer is using GitHub Copilot to generate a function for sorting a large dataset (100,000+ entries). Copilot suggests the following function:

python
Copy
Edit
def sort_data(data):
    return sorted(data)

Why might this function be suboptimal for large datasets?
A) The sorted() function is not efficient for large datasets.
B) The function lacks an explicit sorting algorithm, making it unpredictable.
C) The function does not use an in-place sorting method like data.sort().
D) Copilot might suggest a less efficient sorting algorithm if not prompted explicitly.

4. A developer asks Copilot to generate a function that computes Fibonacci numbers. The suggested function uses recursion but lacks memoization. What is the most likely reason for this omission?

A) Copilot prioritizes standard implementations seen in open-source projects.
B) Copilot avoids generating computationally expensive algorithms.
C) Copilot lacks context about performance optimizations unless explicitly prompted.
D) Copilot assumes the user will manually optimize the generated code.

5. When using GitHub Copilot for SQL query generation, what should developers be cautious about?

A) Copilot always generates the most optimized query.
B) Copilot might suggest queries vulnerable to SQL injection.
C) Copilot does not support SQL query generation.
D) Copilot automatically optimizes database indexes.

6. A team is using Copilot for collaborative development. Developer A enables Copilot in a private repository. Developer B, working in the same repository, notices Copilot's suggestions contain code snippets that look unfamiliar. What might be happening?

A) Copilot is retrieving suggestions from similar open-source projects.
B) Copilot is generating code based on patterns detected in the project’s existing files.
C) Copilot is using external AI models to fetch relevant public code snippets.
D) Copilot is applying prompt engineering techniques to infer missing logic.

7. You request Copilot to generate a SQL query but notice that the suggested query lacks proper indexing considerations. What should you do?

A) Add a comment specifying performance constraints and indexing requirements.
B) Re-prompt Copilot with a more detailed query request.
C) Accept the query and manually optimize it.
D) Adjust Copilot’s model settings to prioritize efficiency.

8. A developer needs Copilot to generate optimized SQL queries. What prompt will yield the best results?

A) # Generate an SQL query to retrieve customer data.
B) # Write an optimized SQL query that retrieves customer data using an index on customer_id.
C) # Write an SQL query for fetching customer data using JOINs.
D) # Generate a basic SQL query to fetch all rows from the customers table.

9. A team is developing a proprietary algorithm and wants to ensure Copilot does not suggest code similar to public implementations. What is the best approach? (Select two)

A) Use private repositories with Copilot to limit exposure to public code patterns.
B) Explicitly describe unique constraints and design principles in inline comments.
C) Turn off Copilot suggestions for functions that contain proprietary logic.
D) Modify Copilot's training dataset to exclude certain open-source projects.

10. According to research on GitHub Copilot's impact, which of the following factors contribute most significantly to increased developer productivity when using Copilot? (Select two correct answers)

A) Reduction in the time spent searching for solutions on external websites.
B) Increased number of lines of code written per hour.
C) A higher frequency of completed pull requests per developer.
D) Faster onboarding of new developers due to AI-assisted code understanding.

11. Which of the following factors most significantly influence GitHub Copilot’s ability to generate high-quality completions? (Select two)

A) The recency of the training data used to build the model.
B) The user's coding patterns and past accepted suggestions.
C) The number of times the user has prompted Copilot in a single session.
D) Whether the user is working in an IDE or a browser-based editor.

12. You are using GitHub Copilot in Visual Studio Code and notice that it frequently suggests outdated API methods that have been deprecated. What are the most effective strategies to improve the relevance of suggestions? (Select two)

A) Add a comment explicitly mentioning the latest API version before the function definition.
B) Use natural language prompts that describe the intent of the code rather than function names.
C) Adjust the settings to restrict Copilot to suggestions from private repositories only.
D) Revert to an older version of GitHub Copilot that was trained on previous API versions.

13. A developer is building a machine learning model in Python. They ask GitHub Copilot to generate a function for model training but notice that Copilot repeatedly suggests suboptimal hyperparameter choices. What is the best way to guide Copilot to generate better completions?

A) Use inline comments to specify preferred hyperparameters and model architectures.
B) Accept a Copilot suggestion and manually tweak the parameters later.
C) Change the function name to include terms like “optimized” or “efficient.”
D) Train Copilot on previous successful models to improve its suggestions.

🚀 🚀 🚀

[paulsuv9]

Lets go! 🚀 🚀 🚀

[Ronaldo-001]

Lezzz go!!

[hemanthnutakki]

Lezzzz goo!!!!!

[hemanthnutakki]

I'm very excited. By understanding developer use cases, AI-driven productivity boosts, and Copilot’s role in SDLC, we can refine our workflows and make the most of its capabilities. We'll also explore Copilot’s limitations and how to leverage the productivity API to measure its impact.

[paulsuv9]

Great enthusiasm! I found that adding specific constraints in the prompt really helps Copilot generate more accurate code. How do you usually refine your prompts? 👍

[johnniekpng]

Go go go...

[Ronaldo-001]

This week was so nice and dug a lot into developer use cases for AI.

Here are my answers for this week:

1. When using GitHub Copilot in a corporate environment, which of the following practices could introduce legal or security risks? (Select two)

• ✅ A) Using Copilot to generate code for encryption algorithms without verification.
• ❌ B) Using Copilot-generated comments to explain complex logic.
• ✅ C) Accepting Copilot-generated code that lacks licensing information.
• ❌ D) Using Copilot to generate documentation for internal APIs.

Explanation:

Copilot-generated encryption algorithms might not be secure or follow industry best practices, leading to security vulnerabilities. Copilot can generate code without proper licensing information, leading to potential legal and intellectual property issues.

---

2. A developer is working with Copilot in Visual Studio and wants to optimize Copilot’s inline chat for debugging. What is the best way to achieve this?

• ✅ A) Rewriting prompts to be more detailed, explicitly mentioning expected outputs.
• ❌ B) Running Copilot in an isolated environment.
• ❌ C) Disabling Copilot's suggestions to avoid incorrect results.
• ❌ D) Avoiding Copilot's inline chat altogether for debugging.

Explanation:

Providing more detailed prompts helps Copilot generate more relevant debugging suggestions.

---

3. A developer is using GitHub Copilot to generate a function for sorting a large dataset (100,000+ entries). Why might this function be suboptimal for large datasets?

def sort_data(data):
    return sorted(data)

• ❌ A) The function uses Python’s built-in sorted() function.
• ❌ B) The function lacks proper error handling.
• ✅ C) The function does not use an in-place sorting method like data.sort().
• ❌ D) The function does not implement a custom sorting algorithm.

Explanation:

sorted(data)creates a new sorted list, increasing memory usage. Usingdata.sort()` sorts the list in place, reducing overhead.

---

4. A developer asks Copilot to generate a function that computes Fibonacci numbers. The suggested function uses recursion but lacks memoization. What is the most likely reason for this omission?

• ❌ A) Copilot avoids recursion in general.
• ❌ B) Copilot assumes performance optimizations are unnecessary.
• ✅ C) Copilot lacks context about performance optimizations unless explicitly prompted.
• ❌ D) Copilot generates inefficient code by default.

Explanation:

Copilot generates code based on common patterns, and unless explicitly prompted, it might not consider optimizations like memoization.

---

5. When using GitHub Copilot for SQL query generation, what should developers be cautious about?

• ❌ A) Copilot might use incorrect SQL syntax.
• ✅ B) Copilot might suggest queries vulnerable to SQL injection.
• ❌ C) Copilot does not support SQL queries.
• ❌ D) Copilot always ensures optimized query performance.

Explanation:

Copilot-generated SQL queries may not use parameterized queries, making them vulnerable to SQL injection.

---

6. A team is using Copilot for collaborative development. Developer A enables Copilot in a private repository. Developer B, working in the same repository, notices Copilot's suggestions contain code snippets that look unfamiliar. What might be happening?

• ❌ A) Copilot is pulling suggestions from external repositories.
• ✅ B) Copilot is generating code based on patterns detected in the project’s existing files.
• ❌ C) Copilot is accessing developer B’s personal GitHub repositories.
• ❌ D) Copilot is using AI-generated randomness.

Explanation:

Copilot suggests code based on the project’s context, including existing files in the repository.

---

7. You request Copilot to generate a SQL query but notice that the suggested query lacks proper indexing considerations. What should you do?

• ✅ A) Add a comment specifying performance constraints and indexing requirements.
• ❌ B) Disable Copilot’s SQL suggestions.
• ❌ C) Accept the generated query and optimize it later.
• ❌ D) Rely on Copilot to always generate optimized queries.

Explanation:

Explicitly defining performance needs in comments improves Copilot’s suggestions.

---

8. A developer needs Copilot to generate optimized SQL queries. What prompt will yield the best results?

• ❌ A) # Write a SQL query to fetch data.
• ✅ B) # Write an optimized SQL query that retrieves customer data using an index on customer_id.
• ❌ C) # Fetch customer data.
• ❌ D) # Generate a basic SQL query.

Explanation:

Providing specific requirements like "optimized" and "using an index on customer_id" ensures Copilot generates a more efficient query.

---

9. A team is developing a proprietary algorithm and wants to ensure Copilot does not suggest code similar to public implementations. What is the best approach? (Select two)

• ✅ A) Use private repositories with Copilot to limit exposure to public code patterns.
• ✅ B) Explicitly describe unique constraints and design principles in inline comments.
• ❌ C) Avoid using Copilot for proprietary code.
• ❌ D) Rely on Copilot’s built-in safeguards.

Explanation:

Private repositories prevent Copilot from using public code patterns in suggestions and providing unique constraints helps Copilot generate more original code.

---

10. According to research on GitHub Copilot's impact, which of the following factors contribute most significantly to increased developer productivity when using Copilot? (Select two)

• ✅ A) Reduction in the time spent searching for solutions on external websites.
• ❌ B) Copilot’s ability to replace all manual coding.
• ❌ C) Copilot’s ability to generate perfect code without human review.
• ✅ D) Faster onboarding of new developers due to AI-assisted code understanding.

Explanation:

Copilot reduces the need to search for external resources, speeding up development.
New developers onboard faster as Copilot assists in understanding and writing code.

---

11. Which of the following factors most significantly influence GitHub Copilot’s ability to generate high-quality completions? (Select two)

• ✅ A) The recency of the training data used to build the model.
• ✅ B) The user's coding patterns and past accepted suggestions.
• ❌ C) Copilot’s randomness.
• ❌ D) The number of programming languages Copilot supports.

Explanation:

More recent training data ensures Copilot generates relevant and up-to-date suggestions.
Copilot learns from user interactions, improving based on accepted completions.

---

12. You are using GitHub Copilot in Visual Studio Code and notice that it frequently suggests outdated API methods that have been deprecated. What are the most effective strategies to improve the relevance of suggestions? (Select two)

• ✅ A) Add a comment explicitly mentioning the latest API version before the function definition.
• ✅ B) Use natural language prompts that describe the intent of the code rather than function names.
• ❌ C) Disable Copilot and write the code manually.
• ❌ D) Ignore the suggestions and try again.

Explanation:

Mentioning the latest API version in comments helps Copilot generate up-to-date suggestions.
Describing the intent ensures Copilot understands the context better and suggests modern alternatives.

---

Also, This week resources are so good ! Thanks @Akash1134 for providing these resources!
Finally excited to share my project soon, which was built with the help of GitHub Copilot.

[SatyamSharma007]

Question: When using GitHub Copilot in a corporate environment, which of the following practices could introduce legal or security risks? (Select two)

✅ A) Using Copilot to generate code for encryption algorithms without verification.

✅ C) Accepting Copilot-generated code that lacks licensing information.

Question: A developer is working with Copilot in Visual Studio and wants to optimize Copilot’s inline chat for debugging. What is the best way to achieve this?

✅ A) Rewriting prompts to be more detailed, explicitly mentioning expected outputs.

Question: A developer is using GitHub Copilot to generate a function for sorting a large dataset (100,000+ entries). Why might this function be suboptimal for large datasets?

✅ C) The function does not use an in-place sorting method like data.sort().

Question: A developer asks Copilot to generate a function that computes Fibonacci numbers. The suggested function uses recursion but lacks memoization. What is the most likely reason for this omission?

✅ C) Copilot lacks context about performance optimizations unless explicitly prompted.

Question: When using GitHub Copilot for SQL query generation, what should developers be cautious about?

✅ B) Copilot might suggest queries vulnerable to SQL injection.

Question: A team is using Copilot for collaborative development. Developer A enables Copilot in a private repository. Developer B, working in the same repository, notices Copilot's suggestions contain code snippets that look unfamiliar. What might be happening?

✅ B) Copilot is generating code based on patterns detected in the project’s existing files.

Question: You request Copilot to generate a SQL query but notice that the suggested query lacks proper indexing considerations. What should you do?

✅ A) Add a comment specifying performance constraints and indexing requirements.

Question: A developer needs Copilot to generate optimized SQL queries. What prompt will yield the best results?

✅ B) # Write an optimized SQL query that retrieves customer data using an index on customer_id.

Question: A team is developing a proprietary algorithm and wants to ensure Copilot does not suggest code similar to public implementations. What is the best approach? (Select two)

✅ A) Use private repositories with Copilot to limit exposure to public code patterns.

✅ B) Explicitly describe unique constraints and design principles in inline comments.

Question: According to research on GitHub Copilot's impact, which of the following factors contribute most significantly to increased developer productivity when using Copilot? (Select two)

✅ A) Reduction in the time spent searching for solutions on external websites.

✅ D) Faster onboarding of new developers due to AI-assisted code understanding.

Question: Which of the following factors most significantly influence GitHub Copilot’s ability to generate high-quality completions? (Select two)

✅ A) The recency of the training data used to build the model.

✅ B) The user's coding patterns and past accepted suggestions.

Question: You are using GitHub Copilot in Visual Studio Code and notice that it frequently suggests outdated API methods that have been deprecated. What are the most effective strategies to improve the relevance of suggestions? (Select two)

✅ A) Add a comment explicitly mentioning the latest API version before the function definition.

✅ B) Use natural language prompts that describe the intent of the code rather than function names.

Question: A developer is building a machine learning model in Python. They ask GitHub Copilot to generate a function for model training but notice that Copilot repeatedly suggests suboptimal hyperparameter choices. What is the best way to guide Copilot to generate better completions?

✅ A) Use inline comments to specify preferred hyperparameters and model architectures.

[paulsuv9]

Great answers! I agree with your points. What do you think are the biggest challenges when using Copilot for large-scale projects? 🚀

[hemanthnutakki]

Great answers!!! I to agree with your points. The way of explanation is too good.

[kalharatennakoon]

Hey @Akash1134 👋
Here are my answers:

1. A & C
2. A
3. C
4. C
5. B
6. B
7. A
8. B
9. A & B
10. A & D
11. A & B
12. A & B
13. A

[paulsuv9]

Nice work! One tip I found useful this week is to use natural language prompts that describe the intent of the code rather than specific function names. It helps Copilot understand the context better. 👍

[Akash1134]

Join the Prompt battle 🤜

💡 Challenge:

Think you can out-prompt GitHub Copilot? Let’s put your prompt engineering skills to the test! Pick a simple coding task and refine your prompt step by step to get the best possible output from Copilot.

How to Participate:

1. Choose a simple task (e.g., "Write a function to reverse a string").
2. Start with a basic prompt and share Copilot's first response.
3. Refine your prompt to make Copilot’s response more efficient, readable, or creative—share the improvements!
4. Bonus: Engage with another participant’s attempt --> can you suggest an even better prompt?

The best one wins the battle ✌️

Example:

• ❌ Basic prompt: "Reverse a string in Python." → Copilot gives a simple function.
• ✅ Improved prompt: "Write a Python function to reverse a string using recursion, with docstrings and edge case handling." → Copilot generates a more optimized version!

Share your Promptness below👇 (we have time till Monday next week)

[paulsuv9]

Challenge accepted! Here’s my attempt:

Initial Prompt:

Refined Prompt 1:

Refined Prompt 2:

Key Differences and Improvements

1. Detailed Docstrings:

• Initial Prompt: The docstring is present but basic.
• Refined Prompt 1: The docstring is similar to the initial prompt.
• Refined Prompt 2: The docstring is more detailed, explaining the parameters, return value, and potential errors.

2. Parameter Validation:

• Initial Prompt: No validation for the type or value of n.
• Refined Prompt 1: No validation for the type or value of n.
• Refined Prompt 2: Includes validation to ensure n is a non-negative integer, raising a ValueError if the input is invalid.

3. Memoization:

• Initial Prompt: Uses a simple list to store the Fibonacci sequence.
• Refined Prompt 1: Introduces memoization by using a list to store previously calculated Fibonacci numbers.
• Refined Prompt 2: Same as Refined Prompt 1, but with added validation and detailed docstrings.

4. Edge Case Handling:

• Initial Prompt: Handles edge cases for n <= 0, n == 1, and n == 2.
• Refined Prompt 1: Similar edge case handling as the initial prompt.
• Refined Prompt 2: Similar edge case handling but with added validation for non-negative integers.

Conclusion
While the initial and refined outputs might look similar at first glance, the refinements add significant value in terms of clarity, robustness, and readability. The detailed docstrings, parameter validation, and memoization technique make the function more reliable and easier to understand.

Looking forward to seeing everyone’s creative prompts! 🚀

[golh30]

Taking up Prompt Battle challenge, below is my prompt engineering demonstration.

1. Initial Prompt 🚧

write a function to find longest palindrom substring in sentence

Respond with function having time complexity O(n^2), used expand_around_center technique. Also Give link about similar code found in gh

2. added system prompt to give context 🐌

You are expert in DSA. write a function to find longest palindrom substring in sentence.

Again respond with O(n^2), but used Dynamic Programming Table

3. improvement in execution time 🎯

You are expert in algorithm to optimize time complexity. write a function to find longest palindrom substring in sentence.

This time copilot picks up best solution out there with O(n), Manacher's Algorithm

✅ Learning:

1. Without context, AI selects a common but non-optimal approach.
2. System prompts influence approach selection.
3. Adding optimization constraints forces AI to choose the best-known solution.

📌 Key Takeaways

1. Default AI Response ≠ Best Solution

• The AI often picks the most common or easy-to-understand method first.

2. Context Matters

• Adding “You are an expert in DSA” shifted the response towards Dynamic Programming.
• The AI interpreted the need for structured and educational content.

3. Optimization Needs Explicit Prompting

• Without asking for the best time complexity, AI might not provide the optimal solution.
• Adding constraints (e.g., “optimize execution time”) improves AI’s response.

Prompt 1

Prompt 2

Prompt 3

[RAVINDRA8008]

Challenge accepted! 🤜🤛 I love how prompt refinement can make Copilot’s responses more efficient and creative. Here’s my attempt:"

➡ Task: Generate Fibonacci numbers
❌ Basic Prompt: "Write a function to generate Fibonacci numbers in Python."
👉 Copilot’s response: Basic iterative function
✅ Improved Prompt: "Write a Python function to generate the first N Fibonacci numbers using recursion. Include a docstring and optimize for efficiency with memoization."
👉 Copilot’s response: Now it includes recursion, memoization, and proper documentation!

"Who’s up for the challenge? Can you refine my prompt even further?" 🚀

This keeps the conversation interactive and invites others to participate. What’s your take—would you like to try a different coding task? 😃

[mvpkenlin]

Init prompt: Could you write a function to do factorial functions in C#?
Refine prompt: How can I optimize this function?

[Sirius-ashwak]

Here are my answers! @Akash1134

1. A & C
2. A
3. C
4. C
5. B
6. A
7. A
8. B
9. A & B
10. A & D
11. A & B
12. A & B
13. A

[paulsuv9]

I’m very excited about this week’s focus on prompt engineering! One tip I found useful is to be very specific with prompts to get the best results from Copilot. Has anyone else noticed a significant improvement in their code quality? 🚀

[paulsuv9]

It might be helpful to add a section on common pitfalls to avoid when using Copilot for prompt engineering.

[hemanthnutakki]

Absolutely! Being specific with prompts makes a huge difference. I've noticed that adding details like expected output format and constraints leads to much better responses from Copilot. It’s like giving clear instructions to a teammate! 🚀🔥

[hemanthnutakki]

Hey @Akash1134
Here are my week 3 answers:
A & C
A
C
C
B
B
A
B
A & B
A & D
A & B
A & B
A

[hemanthnutakki]

Absolutely @paulsuv9 ! Being specific with prompts makes a huge difference. I've noticed that adding details like expected output format and constraints leads to much better responses from Copilot. It’s like giving clear instructions to a teammate! 🚀🔥

[RAVINDRA8008]

Week 3 is looking exciting! 🚀 Prompt engineering is a crucial skill for getting the most out of GitHub Copilot, and it’s great to see a focus on developer productivity and real-world AI use cases.

Some key takeaways for this week:
1️⃣ Mastering prompt crafting – Learning how to frame prompts for optimal Copilot responses.
2️⃣ Exploring developer AI use cases – Seeing how Copilot fits into various workflows.

With all these resources (videos, blogs, and Microsoft Learn modules), there’s plenty to dive into. What part of prompt engineering are you most excited to explore? 🤔

[pintuyadav5468]

answers:

1.Which practices could introduce legal or security risks?
✅ A) Using Copilot to generate code for encryption algorithms without verification.
✅ C) Accepting Copilot-generated code that lacks licensing information.

2.Best way to optimize Copilot’s inline chat for debugging?
✅ A) Rewriting prompts to be more detailed, explicitly mentioning expected outputs.

3.Why is the sorting function suboptimal for large datasets?
✅ C) The function does not use an in-place sorting method like data.sort().

4.Why does Copilot omit memoization in Fibonacci function?
✅ C) Copilot lacks context about performance optimizations unless explicitly prompted.

5.What should developers be cautious about with SQL query generation?
✅ B) Copilot might suggest queries vulnerable to SQL injection.

6.Why do Copilot's suggestions contain unfamiliar snippets in a private repo?
✅ A) Copilot is retrieving suggestions from similar open-source projects.
✅ B) Copilot is generating code based on patterns detected in the project’s existing files.

7.What to do when a Copilot-generated SQL query lacks indexing considerations?
✅ A) Add a comment specifying performance constraints and indexing requirements.

8. Best prompt for optimized SQL query generation?
   ✅ B) # Write an optimized SQL query that retrieves customer data using an index on customer_id.

9. Best approach to prevent Copilot from suggesting public implementations?
   ✅ A) Use private repositories with Copilot to limit exposure to public code patterns.
   ✅ B) Explicitly describe unique constraints and design principles in inline comments.

10.Factors that contribute most to increased developer productivity with Copilot?
✅ A) Reduction in the time spent searching for solutions on external websites.
✅ D) Faster onboarding of new developers due to AI-assisted code understanding.

11.Factors that influence Copilot’s ability to generate high-quality completions?
✅ A) The recency of the training data used to build the model.
✅ B) The user's coding patterns and past accepted suggestions.

12.Strategies to improve Copilot's relevance when it suggests outdated API methods?
✅ A) Add a comment explicitly mentioning the latest API version before the function definition.
✅ B) Use natural language prompts that describe the intent of the code rather than function names.

13.Best way to guide Copilot to generate better ML model training functions?
✅ A) Use inline comments to specify preferred hyperparameters and model architectures.

[johnniekpng]

Here's my answers:

1. Legal and Security Risks (A & C)
   A) Using Copilot to generate code for encryption algorithms without verification: Copilot-generated code should always be reviewed, especially for security-critical components like encryption. Using unverified code could introduce vulnerabilities.
   C) Accepting Copilot-generated code that lacks licensing information: Copilot might generate code snippets from publicly available code that has specific licensing terms. Using such code without understanding and complying with its license could lead to legal issues.

2. Optimizing Copilot Inline Chat for Debugging (A)
   A) Rewriting prompts to be more detailed, explicitly mentioning expected outputs: Providing clear, detailed instructions, including the expected behavior and any error messages, helps Copilot's chat feature understand the context and provide more relevant debugging suggestions.

3. Suboptimal Sorting Function (C)
   C) The function does not use an in-place sorting method like data.sort(): data.sort() modifies the original list directly, which is generally more memory-efficient for large datasets.

4. Recursive Fibonacci without Memoization (C)
   C) Copilot lacks context about performance optimizations unless explicitly prompted: Copilot generates code based on common patterns. The basic recursive Fibonacci implementation is common, but memoization (a performance optimization) requires more specific instruction.

5. SQL Query Generation Cautions (B)
   B) Copilot might suggest queries vulnerable to SQL injection: This is the most critical concern. Copilot may not always generate parameterized queries or properly escape inputs, leaving the application vulnerable to SQL injection attacks.

6. Unfamiliar Code Snippets (B)
   B) Copilot is generating code based on patterns detected in the project’s existing files: Copilot also learns from the code within the current project. If patterns exist in the project's files, Copilot might use those patterns to generate new code, even if those patterns aren't best practices.

7. SQL Query Lacking Indexing (A)
   A) Add a comment specifying performance constraints and indexing requirements: Providing explicit context about indexing helps Copilot generate a more optimized query.

8. Best Prompt for Optimized SQL (B)
   B) # Write an optimized SQL query that retrieves customer data using an index on customer_id: This prompt is the most specific and provides the most relevant information for Copilot to generate an optimized query. It explicitly mentions optimization and the relevant index.

9. Protecting Proprietary Algorithms (A & B)
   A) Use private repositories with Copilot to limit exposure to public code patterns: Using private repositories helps prevent the algorithm from being exposed to the public and reduces the chance of Copilot suggesting similar public implementations.
   B) Explicitly describe unique constraints and design principles in inline comments: Providing detailed comments about the algorithm's unique aspects helps Copilot understand the specific requirements and avoid suggesting generic solutions.

10. Factors Contributing to Increased Productivity (A & D)
    A) Reduction in the time spent searching for solutions on external websites: Copilot can often provide code snippets and solutions directly within the IDE, reducing the need to search online.
    D) Faster onboarding of new developers due to AI-assisted code understanding: Copilot can help new developers understand existing codebases more quickly by providing explanations and generating relevant code examples.

11. Factors Influencing Completion Quality (A & B)
    A) The recency of the training data used to build the model: Copilot's knowledge is based on its training data. More recent data generally leads to better suggestions, especially for newer libraries and frameworks.
    B) The user's coding patterns and past accepted suggestions: Copilot adapts to the user's style and preferences over time, learning from their accepted suggestions and coding habits.

12. Outdated API Suggestions (A & B)
    A) Add a comment explicitly mentioning the latest API version before the function definition: Providing clear context about the desired API version helps Copilot generate relevant suggestions.
    B) Use natural language prompts that describe the intent of the code rather than function names: Describing the purpose of the code, rather than relying on potentially outdated function names, can help Copilot understand the desired functionality and suggest up-to-date methods.

13. Suboptimal Hyperparameter Choices (A)
    A) Use inline comments to specify preferred hyperparameters and model architectures: Providing explicit guidance on the desired hyperparameters and model structure is the most direct way to influence Copilot's suggestions.

[mvpkenlin]

Here are my answers for the week 3

1. When using GitHub Copilot in a corporate environment, which of the following practices could introduce legal or security risks? (Select two)

🟢 A) Using Copilot to generate code for encryption algorithms without verification.
🟢 C) Accepting Copilot-generated code that lacks licensing information.

Note

We always should review and verify the generated code.
if the code lacks licensing information, it could interduce legal or security risks.

2. A developer is working with Copilot in Visual Studio and wants to optimize Copilot’s inline chat for debugging. What is the best way to achieve this?

🟢 A) Rewriting prompts to be more detailed, explicitly mentioning expected outputs.

Note

Only A could help on optimize for debugging

3. A developer is using GitHub Copilot to generate a function for sorting a large dataset (100,000+ entries). Copilot suggests the following function:

python
Copy
Edit
def sort_data(data):
    return sorted(data)

Why might this function be suboptimal for large datasets?
🟢 D) Copilot might suggest a less efficient sorting algorithm if not prompted explicitly.

Note

If we explicitly providing the function is going to sort a very large dataset or even providing what type of data are going to be sorted, Copilot could provide other solutions which could provide better performance.

4. A developer asks Copilot to generate a function that computes Fibonacci numbers. The suggested function uses recursion but lacks memoization. What is the most likely reason for this omission?

🟢 C) Copilot lacks context about performance optimizations unless explicitly prompted.

Note

Same reason as Q3.

5. When using GitHub Copilot for SQL query generation, what should developers be cautious about?

🟢 B) Copilot might suggest queries vulnerable to SQL injection.

Note

This is why Copilot user should always do code review before accept the generated code immediately. It is users' responsibility to verify the security risks.

6. A team is using Copilot for collaborative development. Developer A enables Copilot in a private repository. Developer B, working in the same repository, notices Copilot's suggestions contain code snippets that look unfamiliar. What might be happening?

🟢 B) Copilot is generating code based on patterns detected in the project’s existing files.

Note

Copilot will Learn from patterns: As developers work on a project, GitHub Copilot learns from their coding patterns and preferences, tailoring suggestions accordingly.
Source: Personalized code completion - MS Learn

7. You request Copilot to generate a SQL query but notice that the suggested query lacks proper indexing considerations. What should you do?

🟢 A) Add a comment specifying performance constraints and indexing requirements.

Note

The more contextual information provided, the more fitting the generated code suggestions are.
Source: Provide enough context with details - MS Learn

8. A developer needs Copilot to generate optimized SQL queries. What prompt will yield the best results?

🟢 B) # Write an optimized SQL query that retrieves customer data using an index on customer_id.

Note

A is too simple. C did not provide what column and what kind of join. D fetch all rows.
So correct answer should be B. retrieves customer using index on customer_id

9. A team is developing a proprietary algorithm and wants to ensure Copilot does not suggest code similar to public implementations. What is the best approach? (Select two)

🟢 A) Use private repositories with Copilot to limit exposure to public code patterns.
🟢 B) Explicitly describe unique constraints and design principles in inline comments.

Note

C is incorrect. I tried to go to Copilot setting, i cannot find such setting.
D is incorrect. We cannot list all open-source projects that might have similar proprietary algorithm. And this choice also need custom training.

10. According to research on GitHub Copilot's impact, which of the following factors contribute most significantly to increased developer productivity when using Copilot? (Select two correct answers)

🟢 A) Reduction in the time spent searching for solutions on external websites.
🟢 D) Faster onboarding of new developers due to AI-assisted code understanding.

Note

With copilot, we could minimize context switching. And also can accelerate learning new programming languages and frameworks which helps onboarding of new developers.
Source: Accelerate learning new programming languages and frameworks & Minimizing context switching

11. Which of the following factors most significantly influence GitHub Copilot’s ability to generate high-quality completions? (Select two)

🟢 A) The recency of the training data used to build the model.
🟢 B) The user's coding patterns and past accepted suggestions.

Note

More recent training data could allow GitHub Copilot to generate relevant and up-to-date generations. And User's codng patterns and past accept suggestions are part of the GitHub could generate Personalized code completion.

12. You are using GitHub Copilot in Visual Studio Code and notice that it frequently suggests outdated API methods that have been deprecated. What are the most effective strategies to improve the relevance of suggestions? (Select two)

🟢 A) Add a comment explicitly mentioning the latest API version before the function definition.
🟢 B) Use natural language prompts that describe the intent of the code rather than function names.

Note

C is incorrect because it does not help to filter out the outdated API.
D is incorrect, it also does not help to filter out the outdated API.

13. A developer is building a machine learning model in Python. They ask GitHub Copilot to generate a function for model training but notice that Copilot repeatedly suggests suboptimal hyperparameter choices. What is the best way to guide Copilot to generate better completions?

🟢 A) Use inline comments to specify preferred hyperparameters and model architectures.

Note

B is incorrect, although it works but it is not the suggested way when using Copilot.
C is incorrect, it does not help.
D is incorrect, train Copilot on previous successful models does not help.

[mvpkenlin]

@Akash1134
I found a mistake link in your first post in this thread, "Implement code improvements using GitHub Copilot tools".

Currently it is linking to "https://learn.microsoft.com/en-us/training/modules/developer-use-cases-for-ai-with-github-copilot/"

And it should be "https://learn.microsoft.com/en-us/training/modules/implement-code-improvements-using-github-copilot-tools/"

[mvark]

My entry for the GitHubCopilotChallenge - NutriScan, a minimalist web application that turns your mobile browser into a barcode scanner and tells you whether a packaged food item is Ultra-Processed or not. No dedicated smartphone app is required unlike other scanning apps; it runs with just HTML and JavaScript!

JavaScript has changed significantly in the last decade. Many of the features that were part of helper libraries like jQuery are now included natively. As I wanted the app to load very fast, GitHub Copilot generated lean code and GitHub Copilot Chat explained the code where I needed help in understanding.

Overview
NutriScan allows you to scan a product barcode using your smartphone's camera and then redirects you to the corresponding product's page on Open Food Facts to display nutrition information.

Open Food Facts is the world's largest open source food database, a sort of Wikipedia of Food. Nutrition information for more than 3.6 million products is available on the OFF website. You can access this data through their website, API, or as a data dump.

Open Food Facts employs a nutritional rating system that rewards products rich in beneficial nutrients like fiber and protein, while penalizing those high in less desirable components such as fats, sugars, and sodium.

I'm an active contributor to the Open Food Facts database and I highly encourage like-minded people to add missing data.

My sample code scans an EAN-13 barcode from a packaged food product and then fetches its:

• Name
• Image
• Nutri-Score giving an indication of the nutritional quality of a food item
• Nova classification assigning foods a score from 1 to 4 — where 1 represents unprocessed foods and 4 denotes ultra-processed foods
• using the Open Food Facts REST API.

How to Use

• Access the Web App: Open the NutriScan web page in your mobile browser.
  

• Scan Barcode: The app will request access to your device's camera. Grant permission and scan the barcode of a packaged food product.

• View Information: The app will display the product's name, image, Nutri-Score, and Nova grouping.
  

Key Features

• Barcode Scanning: Utilises Minhaz's open-source HTML5 QR Code & Barcode Scanner JavaScript library to access a mobile device's camera and scan barcodes directly within the browser.
• Open Food Facts Integration: Redirects to Open Food Facts to display detailed nutrition information for the scanned product or fetches product details via the Open Food Facts REST API.
• Minimalist Design: The code is deliberately kept lean to demonstrate the minimum amount of code required for a functional app.
• Low Data Transfer: The total data transferred to and from the app is less than 400kb.
• No App Required: Eliminates the need for users to install a dedicated barcode scanner app.

Code Structure
The repository includes:

• HTML file with the basic structure of the web page. It includes JavaScript code to handle camera access, barcode scanning, and data retrieval from Open Food Facts.
• The code for the JavaScript file html5-qrcode.min.js can be copied from Minhaz's GitHub repository, which hosts the HTML5 QR Code & Barcode Scanner JavaScript library.

Troubleshooting
During development I discovered that there were instances where nova_group and nutriscore_grade fields in the JSON were missing from the API response. This was causing an error. With the help of GitHub Copilot & remote debugging through the Android device, I identified & fixed the issue.

I've also added "Thing Near Me" the other app I co-created with Copilot to the repo WebApps. I plan to add more code samples as I continue exploring interesting public REST APIs.

[Parthapnath]

Answers:
1.A, C
2.A
3.C
4.C
5.B
6.B
7.A
8.B
9.A, B
10.A, D
11. A, B
12.A, B
13.A

[Parthapnath]

Week 3 questions were a bit tricky but it was also fun learning about prompt engeneering and productivity.