ByteDance·Data Scientist·Technical Phone Screen
Jun 2026
New-grad ML/data scientist loop at ByteDance. Six questions back to back, all technical, no fluff. The range was pretty wide: systems, retrieval, classic ML theory, and LLM alignment stuff. Left feeling like I'd studied the wrong half of the syllabus.
- You need to deploy a multimodal model (text plus image, video, or audio) under tight GPU memory, latency, and cost constraints. How would you redesign both the model and the serving infrastructure to hit those constraints without wrecking quality?
- You have a large video corpus where each video already has a caption and precomputed embeddings. How would you build a retrieval system that returns relevant results fast and with high recall?
- What is overfitting, how do you detect it, and what are the most effective ways to reduce it in deep learning?
- Explain the intuition behind dropout and why the inverted dropout trick keeps activation magnitudes consistent between training and inference. When does dropout hurt more than it helps?
- Compare Batch Normalization, Layer Normalization, Group Normalization, Instance Normalization, and RMSNorm. What statistics does each use, how do they behave differently at training vs inference, and why do transformers tend to prefer certain ones?
- Walk through how reinforcement learning is used to fine-tune LLMs after supervised training. Cover the reward model, KL regularization, common failure modes, and how you'd evaluate the result. How does PPO-based alignment compare to direct preference optimization approaches?
“This one sprawled in every direction and I didn't manage it well.” The rest of the author's notes on Data Scientist interview at ByteDance, Technical Phone Screen round, covers how they worked through the question, what the panel pushed back on, and what they would do differently.
View Post