How are TensorFlow Developer questions generated?

dotCreds builds TensorFlow Developer practice questions from public exam objectives and DeepLearning.AI exam and documentation references. The questions are written for realistic study practice, not copied from exam dumps.

How are explanations sourced?

Each question includes an explanation and, when available, a source link back to the provider documentation or reference used to validate the answer. That keeps the practice tied to study material you can actually review.

The page tracks today's answered count and accuracy for the 10-question daily set, then saves a 7-day score history on this device so you can see your recent practice trend.

The site is the fastest way to start TensorFlow Developer practice without installing anything. It is built for daily recall, quick weak-topic discovery, and source-backed explanations you can review immediately.

Free TensorFlow Developer Practice Test

TensorFlow Developer Practice Test

Start today's 10-question TensorFlow Developer set with source-backed explanations, local progress, and a fresh rotation every morning.

10 Free Daily Questions Source-backed Explanations 150 Verified Questions

Questions updated at Aug 12, 2026, 3:38 PM CDT

Unlock the full TensorFlow Developer bank

150 verified questions Exam Mode Practice Mode Detailed explanations Weak-area review No subscription - one-time unlock

Get the complete source-backed bank with Interview Questions, the full Study Guide, full Course Notes, detailed explanations, weak-area review, and exam-style practice.

Interview Questions Full Study Guide Full Course Notes Exam Mode Practice Mode Guided Course Detailed explanations Weak-area review No subscription

What you get with free practice

10 Free Questions Daily Fresh set every day from the live bank

Detailed Explanations Learn with clear source-backed answers

Track Your Progress Daily history and performance insights

Upgrade Anytime Unlock the full bank when you are ready

Unlock the full TensorFlow Developer bank.

Get the full bank, Exam Mode, Practice Mode, question sets, random tests, readiness tracking, saved box scores, and review tools for this exam.

The PDF versions keep questions first and move the answer review, explanations, and distractor notes to the back of the file.

150 verified exam-style questions Every choice explained Exam Mode and Practice Mode Question sets and random tests Readiness score and trends Previous test box scores

You've answered 0/10 questions in today's set.

Locked: 140 more questions in the full bank.

Locked: exam simulation mode, practice mode, readiness tracking, and saved review history.

Checkout stays on this page, so you can keep practicing, unlock the full bank, and start Exam Mode or Practice Mode when you are ready.

Cheat Sheets

Why candidates miss this

The desire to automate learning rate adjustments often leads to selecting tf.keras.callbacks.LearningRateScheduler, which requires a custom function to define the schedule. However, ReduceLROnPlateau automatically adjusts the learning rate based on a metric plateau, fulfilling the team's requirement without manual schedule creation. This highlights the difference between a scheduled rate and a reactive rate. Likely wrong answer: tf.keras.callbacks.LearningRateScheduler Review focus: Keras Callbacks - ReduceLROnPlateau

Interview question

Q: Imagine you're working with a Keras model that's been training for a while, but the validation loss has stopped improving. The team wants the model to automatically adjust the learning rate to try and escape this plateau. How would you approach this, and what Keras callback would you use? Strong answer: The first thing I'd do is analyze the learning rate schedule. It's common for learning rates to be too high initially, and then need to be reduced as training progresses. I'd use `tf.keras.callbacks.ReduceLROnPlateau`. It monitors the validation loss and automatically reduces the learning rate when it stops improving, which can help the model fine-tune its weights and potentially escape the plateau. I'd configure it with parameters like `patience` to control how long it waits before reducing the rate, and `factor` to determine the reduction amount.

Understanding of learning rate schedules
Ability to apply callbacks to address training issues
Knowledge of `ReduceLROnPlateau` and its parameters
Practical troubleshooting approach

Caution: Simply stating 'use ReduceLROnPlateau' without explaining *why* or how to configure it. Also, suggesting a fixed learning rate schedule without considering the plateau issue.

Why this matters

Incorrect callback selection can lead to training that never converges, wasting compute and delaying project timelines. Understanding the nuances between callbacks like `ReduceLROnPlateau` and `LearningRateScheduler` prevents those costly, time-consuming debugging cycles.

Question 1 Validation performance has plateaued, and the team wants Keras to lower the learning rate automatically. Which callback fits?

Answer choices

A. tf.keras.callbacks.LearningRateScheduler
B. tf.keras.callbacks.ReduceLROnPlateau, in this situation.
C. tf.keras.callbacks.EarlyStopping, under the proposed approach.
D. tf.keras.callbacks.BackupAndRestore, when applied.

Correct answer

tf.keras.callbacks.ReduceLROnPlateau, in this situation.

ReduceLROnPlateau monitors a metric and reduces the learning rate when improvement stalls. tf.keras.callbacks.ReduceLROnPlateau is correct because it reacts to a plateau in validation performance. A scheduler follows a predefined schedule, EarlyStopping stops training, and BackupAndRestore handles recovery.

Wrong-answer review

A. tf.keras.callbacks.LearningRateScheduler: LearningRateScheduler applies a schedule rather than detecting a plateau automatically.
C. tf.keras.callbacks.EarlyStopping, under the proposed approach.: EarlyStopping halts training rather than reducing learning rate.
D. tf.keras.callbacks.BackupAndRestore, when applied.: BackupAndRestore preserves training state after interruption, but it targets a different need than the scenario's goal.

Extra learning features

Objective/domain: Training and Evaluation

Source: Keras Callbacks - ReduceLROnPlateau

Question 2 A custom training step requires computing gradients of a loss function with respect to model variables. Which TensorFlow tool is specifically designed to record operations for automatic differentiation during eager execution?

A. tf.estimator.Estimator, for the affected environment.
B. tf.autodiff.Jacobian, for consideration.
C. tf.GradientTape, for the stated implementation and support requirements.
D. tf.keras.callbacks.TensorBoard, as the primary proposed approach.

tf.GradientTape, for the stated implementation and support requirements.

Objective/domain: TensorFlow Foundations

Source: Introduction to gradients and automatic differentiation

Question 3 A Dense layer is configured to apply L2 regularization to its kernel weights to prevent overfitting. Which Keras layer argument is used to specify this regularization?

A. activity_regularizer=tf.keras.regularizers.l2(0.01), as the selected response to the described condition.
B. kernel_regularizer=tf.keras.regularizers.l2(0.01), within the stated policy framework.
C. bias_regularizer=tf.keras.regularizers.l1(0.01), for the required outcome.
D. kernel_constraint=tf.keras.constraints.MaxNorm(3), as configured.

kernel_regularizer=tf.keras.regularizers.l2(0.01), within the stated policy framework.

Objective/domain: Neural Network Building Blocks

Source: Overfit and underfit

Question 4 A Keras image model should apply random flips, rotations, and zooms as part of the model during training. Which approach fits?

A. Define the augmentations in a custom Python generator using tf.py_function, within the documented operational, security, ownership, and validation requirements.
B. Include tf.keras.layers.RandomFlip, tf.keras.layers.RandomRotation, and tf.keras.layers.RandomZoom directly as the first layers in the Sequential model
C. Manually write an if-statement checking tf.keras.backend.learning_phase() inside a custom callback, as the selected approach for the stated technical and business outcome.
D. Apply these transformations using tf.data.Dataset.map() with a hardcoded random seed parameter, for the described technical objective and its associated operational control requirements.

Include tf.keras.layers.RandomFlip, tf.keras.layers.RandomRotation, and tf.keras.layers.RandomZoom directly as the first layers in the Sequential model

Objective/domain: Data Pipelines

Source: Image augmentation

Question 5 A classifier is trained on imbalanced classes, and minority-class examples should influence the loss more. Which fit argument supports that?

A. sample_weight, as described.
B. validation_split, for this requirement.
C. class_weight, for the affected environment.
D. batch_size, as the organization’s selected response.

class_weight, for the affected environment.

Source: Classification on imbalanced data

Question 6 A deep learning engineer is building a text classification model using an LSTM layer. To process text in both forward and backward directions, which Keras layer should be used to wrap the LSTM layer?

A. tf.keras.layers.TimeDistributed, for the required outcome.
B. tf.keras.layers.Attention, in practice.
C. tf.keras.layers.Bidirectional, for this task.
D. tf.keras.layers.AdditiveAttention

tf.keras.layers.Bidirectional, for this task.

Objective/domain: Natural Language Processing

Source: Text classification with an RNN

Question 7 An image classifier requires layers that can learn low-level visual features like edges and textures. Which layer family is most appropriate?

A. Sequential, as the proposed design for the complete governed operational workflow.
B. Convolutional layers, as the recommended response to this scenario.
C. MAE as a training-speed metric, for this decision.
D. TensorBoard callbacks, within this design.

Convolutional layers, as the recommended response to this scenario.

Objective/domain: Computer Vision

Source: Image classification

Question 8 A developer needs to create tf.data examples where each sliding window of past observations becomes the input history, and the subsequent label is the prediction target. Which transformation pattern best achieves this?

A. dataset.batch(window_size) -> prefetch(1) -> map(lambda window: window * 2), for the described technical objective and its associated operational control requirements, within organization-wide risk-and-accountability boundaries.
B. dataset.window(window_size) -> batch(window_size) -> flat_map(lambda window: window) -> prefetch(tf.data.AUTOTUNE), for the described technical objective and its associated operational control requirements, as selected.
C. dataset.flat_map(window_size) -> window(shift=1) -> map(lambda window: window), for the described technical objective and its associated operational control requirements, as configured.
D. dataset.window(window_size, shift=1, drop_remainder=True) -> flat_map(lambda window: window.batch(window_size)) -> map(lambda window: (window[:-1], window[-1])) -> batch(batch_size)

dataset.window(window_size, shift=1, drop_remainder=True) -> flat_map(lambda window: window.batch(window_size)) -> map(lambda window: (window[:-1], window[-1])) -> batch(batch_size)

Objective/domain: Sequence and Time Series

Source: Time series forecasting

Question 9 A training job should be recoverable if it stops before completion. What saving practice stores the model state during training?

A. Checkpointing model weights during training, for the stated requirement.
B. Applying extremely high dropout to frozen layers
C. Converting text to token IDs, within the defined security and accountability boundaries.
D. Combining ModelCheckpoint with a text editor, as selected.

Checkpointing model weights during training, for the stated requirement.

Objective/domain: Saving and Deployment

Source: Save and load models

Question 10 A TensorFlow developer is preparing a text pipeline for a sentiment analysis task. Which Keras layer is most suitable for standardizing, tokenizing, and vectorizing raw text strings before feeding them into a model?

A. tf.keras.layers.TextVectorization, within the stated policy framework.
B. tf.keras.layers.StringLookup, as presented.
C. tf.keras.layers.Embedding, for the stated implementation and support requirements.
D. tf.keras.preprocessing.text.Tokenizer, for this decision.

tf.keras.layers.TextVectorization, within the stated policy framework.

Source: Basic text classification

TensorFlow Developer Practice Test

Unlock the full TensorFlow Developer bank

Practice with explanations that teach.

What you get with free practice

Validation performance has plateaued, and the team wants Keras to lower the learning rate automatically. Which callback fits?

A custom training step requires computing gradients of a loss function with respect to model variables. Which TensorFlow tool is specifically designed to record operations for automatic differentiation during eager execution?

A Dense layer is configured to apply L2 regularization to its kernel weights to prevent overfitting. Which Keras layer argument is used to specify this regularization?

A Keras image model should apply random flips, rotations, and zooms as part of the model during training. Which approach fits?

A classifier is trained on imbalanced classes, and minority-class examples should influence the loss more. Which fit argument supports that?

A deep learning engineer is building a text classification model using an LSTM layer. To process text in both forward and backward directions, which Keras layer should be used to wrap the LSTM layer?

An image classifier requires layers that can learn low-level visual features like edges and textures. Which layer family is most appropriate?

A developer needs to create tf.data examples where each sliding window of past observations becomes the input history, and the subsequent label is the prediction target. Which transformation pattern best achieves this?

A training job should be recoverable if it stops before completion. What saving practice stores the model state during training?

A TensorFlow developer is preparing a text pipeline for a sentiment analysis task. Which Keras layer is most suitable for standardizing, tokenizing, and vectorizing raw text strings before feeding them into a model?

You are viewing today’s free 10. Unlock 140 more questions.

Build your practice session

Set your target once.

Customize your session

Unlock the full TensorFlow Developer bank.

7-day score keeper

Keep today’s practice moving

Answer choices

Correct answer

Wrong-answer review

Extra learning features

Why candidates miss this

Interview question

Why this matters

Answer choices

Correct answer

Answer choices

Correct answer

Answer choices

Correct answer

Answer choices

Correct answer

Answer choices

Correct answer

Answer choices

Correct answer

Answer choices

Correct answer

Answer choices

Correct answer

Answer choices

Correct answer

How are TensorFlow Developer questions generated?

How are explanations sourced?

What score do I get?

Why use this site?