Gradient Descent from Scratch | Dataford Interview Questions

Implement gradient descent for a simple function from scratch in Python. Asked in the Phone Screen stage. This ML coding question was part of a 1-hour phone screen and was expected to take about 20 minutes.

Contract

Implement gradient_descent(initial_x, learning_rate, iterations) for f(x) = (x - 4)^2. Its derivative is 2 * (x - 4). Return the final x after exactly iterations updates using x = x - learning_rate * derivative(x). Return the result rounded to 10 decimal places.

Constraints

-10^9 <= initial_x <= 10^9
0 < learning_rate <= 0.5
0 <= iterations <= 10^5
The objective function is f(x) = (x - 4)^2
Return the final value rounded to 10 decimal places

Function Signature

def gradient_descent(initial_x, learning_rate, iterations):

Your question is Gradient Descent from Scratch. Start with the requirements in the Question tab.

Run and submit as often as you like. When you're ready, talk me through your approach or go straight to the code.

Problem

Contract

Constraints

-10^9 <= initial_x <= 10^9
0 < learning_rate <= 0.5
0 <= iterations <= 10^5
The objective function is f(x) = (x - 4)^2
Return the final value rounded to 10 decimal places

Function Signature

def gradient_descent(initial_x, learning_rate, iterations):

Your question is Gradient Descent from Scratch. Start with the requirements in the Question tab.

Run and submit as often as you like. When you're ready, talk me through your approach or go straight to the code.