TikTok Interview Questions — Page 21

TikTok·Software Engineer·Technical Phone Screen

Jun 2026

TikTok Data Engineer interview with a SQL-heavy technical screen. The question was pretty involved and required walking through a query step by step rather than just writing one from scratch.

Given two tables and a specific SQL query, explain exactly what the result set looks like: which rows get returned, what each column contains, how joins, filters, aggregations, and window functions shape the output, how NULLs and duplicates are handled, and what the final ordering is.

“This tripped me up more than I expected.”

View Post

TikTok·Data Scientist·Technical Phone Screen

Jun 2026

TikTok data scientist interview that went deep on XGBoost internals, hyperparameter tuning, and a pretty gnarly imbalanced classification scenario. The questions felt more like a written exam than a conversation, and the level of detail expected was honestly a lot for a single session.

Walk through how XGBoost's tree booster works: what objective does it optimize, how does the second-order Taylor approximation produce a split gain formula, and what roles do lambda, alpha, gamma, and eta each play in that gain and in pruning decisions?
For tabular classification with XGBoost, describe the most impactful hyperparameters, including max_depth, max_leaves, min_child_weight, subsample, colsample_bytree and colsample_bylevel, eta, n_estimators, lambda, alpha, gamma, max_delta_step, scale_pos_weight, and monotone_constraints. For each, explain how it shifts bias versus variance and what it does to training time.
You're building a model to detect bad sellers where the positive rate is around 0.5%. Design a full tuning plan that addresses data splitting to avoid leakage, choosing the right offline metric, setting an operating threshold from a cost matrix, applying early stopping properly, and what plots or diagnostics you'd use to catch overfitting and leakage.
After training the bad-seller model, how would you calibrate the output probabilities, explain model decisions to investigators using something like SHAP, and prevent adversaries from reverse-engineering the detection rules?

“This is the kind of question where you think you know XGBoost until someone asks you to derive it.”

View Post

TikTok·Machine Learning Engineer·Technical Phone Screen

Jun 2026

TikTok ML engineer interview with a coding question on matrix traversal. Nothing too surprising but the spiral order detail trips people up if you're not careful about boundary conditions.

Given an m x n matrix, return all elements in clockwise spiral order.

“I've seen this before so I wasn't panicking, but I still fumbled the boundary shrinking logic midway through.”

View Post

TikTok·Machine Learning Engineer·Onsite - Multi Round

Jun 2026

TikTok ML Engineer onsite, deep resume drill on a VLM project. The whole session was basically one long question where they just kept pulling on threads until something frayed.

Walk me through the VLM project on your resume. What problem were you solving and how did you define success?
Describe the architecture: vision encoder, language model, and how you fused the two modalities.
What was frozen versus trainable during your training runs, and why did you make those choices?
Tell me about your training data: sources, label types, known biases, and how you handled train/val/test splits.
Walk through your full training recipe: objectives, stages, key hyperparameters, and how you evaluated progress.
Was the model trained end-to-end or modular? What were the practical trade-offs of that choice?
Where does inference time actually go in your system, and what optimizations did you try?
What are the known failure modes of your model, and what concrete steps would you take to address them?

“I started with the task framing and metrics, which felt natural, but I stumbled when they asked how offline metrics mapped to actual product impact.”

View Post

TikTok·Data Scientist·Onsite - Behavioral / Leadership

Jun 2026

Behavioral round at TikTok for a Data Scientist role, focused on how you handle friction with stakeholders during a project. Pretty standard stuff but the follow-up probing made it trickier than expected.

Tell me about a time you had a conflict with stakeholders during a project. What did you do and how did it turn out?

“I had a story ready but stumbled when they pushed on what I personally did versus what the team did.”

View Post

TikTok·Machine Learning Engineer·Technical Phone Screen

Jun 2026

TikTok ML engineer round with a dynamic programming twist I didn't fully see coming. The problem looked like a standard word break at first glance but the second-smallest cost angle made it genuinely harder.

Given a string and a dictionary of words, a valid segmentation splits the string into dictionary words whose concatenation equals the original. The cost is the number of words used. Return the second-smallest cost across all valid segmentations, or -1 if fewer than two valid segmentations exist. Also explain your algorithm's time and space complexity and how you'd handle large inputs.

“My first instinct was to track just the minimum cost per index, which is basically the standard word break DP.”

View Post

TikTok·Software Engineer·Onsite - Coding / Algorithms

Jun 2026

Four-question algorithmic gauntlet for a TikTok SWE role. The problems ranged from classic DP to some greedy reasoning and a sneaky simulation at the end. Felt like a lot to cover in one sitting.

Implement a wildcard pattern matcher where '?' matches any single character and '*' matches any sequence including empty. Discuss DP versus greedy approaches and analyze the complexity of each.
Design a data structure initialized with two integer arrays A and B that supports an update operation adding a value to B[i], and a query operation counting pairs where a + b equals a given total. Handle up to 100,000 updates and queries efficiently.
Implement an LRU cache supporting get and put in O(1) time. Describe your design choices.
Given arrays of distances and speeds for n threats, starting at time 0 you may neutralize one threat per minute. A threat reaches the base when elapsed time is at least dist[i]/speed[i]. If any threat has already arrived before you act at a given minute, the process ends. Return the maximum number of threats you can neutralize.

“I went straight to 2D DP which felt safe, but they pushed back asking if I could do better on space.”

View Post

TikTok·Data Scientist·Technical Phone Screen

Jun 2026

Product-facing data science round at TikTok that went deep on tree-based ensembles. The whole session was basically one long question about Random Forests vs gradient boosting, with follow-ups that kept branching into production concerns and then a curveball about feature scaling. More technical than I expected for a product DS role.

Compare Random Forests and Gradient-Boosted Decision Trees like XGBoost across bias/variance, interpretability, training and inference speed, and robustness to overfitting. How does the difference between bagging and sequential boosting explain each of those trade-offs?
In a production environment, when would you choose a Random Forest over a gradient-boosted model or vice versa? Think about accuracy ceiling, tuning complexity, latency, noise robustness, calibration, and what happens when the data distribution shifts.
Do tree-based models need feature standardization or normalization? Walk through the theoretical reason and mention any practical cases where it might still matter.

“This is where I spent most of the interview and also where I fumbled the most.”

View Post

TikTok·Software Engineer·Onsite - Behavioral / Leadership

Jun 2026

Behavioral screen for a Software Engineer role at TikTok. Pretty standard stuff: background, where you're from, and one deep-dive into a resume project. Nothing technically complex but the project question had more layers than I expected.

Walk us through your background: education, current role, and what you spend most of your time on.
Where are you from, and does that affect anything like your time zone or how you prefer to work?
Pick a project from your resume and explain the business context, what you personally owned, which features you focused on, and what the hardest problem was along with your measurable impact.

“Felt fine until I realized I was rambling about my education for way too long and barely got to what I actually do now.”

View Post

TikTok·Data Scientist·Technical Phone Screen

Jun 2026

TikTok data scientist interview with a causal inference question that felt more like a mini research design exercise than a typical DS screen. The question had real depth to it and pushed me to think carefully about selection bias in a product context.

Management wants to understand how negative customer reviews affect whether a user buys the same coupon again. How would you design an analysis to measure this effect, and what methods would you use to separate the review's impact from other confounding factors?

“This one took me a second to frame properly.”

View Post