Senior ML Systems Engineer interview at Meta
The FP8 Quantization Tax
“You moved your matmuls to FP8 on H100s expecting a 2× speedup. You got 25%. Is FP8 overhyped, or did you do it wrong?”
LLM System Design #58
Sep 2, 2026
Meta interview traps
42 traps set in interviews at Meta. Each one: the interviewer's question, the answer most candidates give, and the mechanism that breaks it.
Roles: Senior ML Engineer (14), Senior AI Engineer (12), Senior Computer Vision Engineer (4), AI Engineer (2), Computer Vision Engineer (1).
From Machine Learning System Design Interview, Advanced Deep Learning Interview Questions, LLM System Design Interview, Advanced NLP Interview Questions, Computer Vision Interview Questions, LLM Inference Interview Questions and RAG Interview Questions.
Each trap is set in an interview at Meta. AI Interview Prep isn't affiliated with Meta.
Senior LLM Research Engineer interview at Meta
The SwiGLU Capacity Trap
“You swapped your ReLU FFN for SwiGLU, loss dropped 2%, and you shipped it. Why should I reject that result?”
LLM System Design #56
Aug 31, 2026Senior ML Engineer interview at Meta
The Redundant Tool Paradox
“You’re building synthetic tool-call training data with Toolformer’s filter, keep the call if the tool output raises the likelihood of the correct continuation. Your eval improves. Production accuracy doesn’t move. What’s wrong with the filter?”
LLM Inference #11
Aug 9, 2026Senior AI Engineer interview at Meta
The Zombie Hub Trap
“Your HNSW index takes a delete every few seconds in production. Do you remove the node and repair the graph?”
RAG #11
Jul 15, 2026Senior LLM Engineer interview at Meta
The Low-Rank Capacity Trap
“Your team used a standard LoRA for continual pre-training on a massive, complex new domain to save vRAM. But the model’s performance is severely inferior to a full fine-tune. How do you modify the adapter architecture and optimization strategy to bridge this gap?”
ML System Design #48
Jun 5, 2026Senior Computer Vision Engineer interview at Meta
The Invariance Illusion
“Your medical imaging model shows a flawless 0.99 AUC offline, but plummets to 0.65 when deployed because real-world clinic scans undergo minor 3-degree rotations and arbitrary cropping. How do you redesign your CI/CD evaluation pipeline to catch these semantic vulnerabilities automatically before hitting runtime?”
ML System Design #44
Jun 1, 2026Senior Staff AI Engineer interview at Meta
The Overfitting Illusion
“Before spinning up a massive distributed training run across a cluster of 512 H100 GPUs, you mandate that your team run a test to deliberately overfit the architecture on a single batch of data. Your team objects, arguing that compute is too expensive to waste on memorizing one batch. How do you justify this?”
ML System Design #43
May 31, 2026Senior ML Engineer interview at Meta
The Base-Rate F1 Trap
“An engineer shows you a binary classification model boasting a phenomenal 0.90 F1-score on a newly curated validation set, claiming it’s ready for production deployment. Before even looking at the architecture, you flag this metric as a potential illusion. What hidden data profile characteristic are you suspecting, and how do you prove it?”
ML System Design #42
May 30, 2026Senior AI Engineer interview at Meta
The Uncertainty Loop Paradox
“We have a 10-million sample unlabelled dataset and want to fine-tune a Llama-3 70B model while minimizing manual annotation costs. How do you design an Active Learning loop to selectively label the most uncertain points?”
ML System Design #37
May 25, 2026Senior ML Engineer interview at Meta
The Weighted Cross-Entropy Trap
“You’re training a fraud detection model on an extremely imbalanced production stream, 1 fraud sample for every 10,000 legitimate transactions. How do you construct the loss function to ensure the model actually learns the rare class without collapsing?”
ML System Design #35
May 23, 2026Senior ML Ops Engineer interview at Meta
The Streaming Bias Trap
“You need to uniformly sample an unbounded, real-time stream of raw user events into a fixed-size memory buffer for downstream training. The total volume is unknown and cannot fit in RAM. What specific algorithm guarantees a mathematically uniform distribution, and what happens to your model’s bias if you skip it?”
ML System Design #33
May 21, 2026Senior ML Engineer interview at Meta
The Transformation Debt Trap
“We need to ingest petabytes of raw, unstructured data, text, images, and audio. for our new multimodal GenAI pipeline. Everyone loves the modern data stack, so should we use ELT to dump it all into the data lakehouse as fast as possible and transform it later?”
ML System Design #30
May 18, 2026Senior ML Engineer interview at Meta
The Correlational Trap
“Your consumer app’s new recommendation model increased short-term engagement metrics (like “time spent”
ML System Design #29
May 17, 2026Senior ML Engineer interview at Meta
The Clickbait Trap
“You built a recommendation engine for an e-commerce site. Precision, recall, and CTR are all hitting all-time highs, but the actual user sign-up rate hasn’t budged. What fundamental assumption in your proxy metrics just failed in production, and how do you course-correct?”
ML System Design #27
May 15, 2026Senior ML Engineer interview at Meta
The Inference Bottleneck Illusion
“You’ve built a two-tower recommendation system balancing high recall and high precision. The problem? It takes 400ms to run the pipeline, but product demands a strict 100ms SLA. Where do you cut latency without destroying the user experience?”
ML System Design #26
May 14, 2026Senior AI Engineer interview at Meta
The FP32 Hidden Tax
“You load a 7-billion parameter model onto an 80GB A100 in BF16. You calculate the weights take up a mere 14 gigabytes. But the moment you initialize your Adam optimizer and take a single training step, the script violently crashes with an Out-Of-Memory (OOM) error. Down to the exact byte multipliers, what hidden variables just silently consumed the vast majority of your memory footprint?”
LLM System Design #45
May 8, 2026Senior AI Engineer interview at Meta
The Isomorphic MLP Trick
“You are upgrading our legacy ReLU-based transformer to use modern SwiGLU activations. To keep the architecture consistent, you leave the feed-forward dimension up-projection ratio at the standard 𝟒 × 𝐝_𝐦𝐨𝐝𝐞𝐥. What subtle but massive architectural mistake have you just made regarding your parameter budget?”
LLM System Design #36
Apr 29, 2026Senior AI Engineer interview at Meta
The Normalization Paradox
“We are optimizing our next 70B parameter LLM and decide to swap standard LayerNorm for RMSNorm. Why are we doing this?”
LLM System Design #34
Apr 27, 2026Senior PyTorch Engineer interview at Meta
The AdamW Memory Trap
“Your 70B parameter model training job on 1,024 H100s just crashed on day 5. You successfully load the saved model.state_dict() and resume, but your loss immediately spikes to the moon, destroying weeks of progress. What happened?”
LLM System Design #32
Apr 22, 2026Senior AI Engineer interview at Meta
The Precision Allocation Trap
“You’re trying to fit a 40B parameter model on 8 H100s. To save memory, you cast the entire model and optimizer state to BF16. Your training instantly goes haywire and diverges. What critical mixed-precision rule did you just violate, and why?”
LLM System Design #30
Apr 20, 2026Principal AI Engineer interview at Meta
The Compute-Without-Data Trap
“We just secured a cluster of 100,000 H100s, but we’ve completely run out of high-quality internet text. How does entering this ‘data-constrained’ regime completely invert our standard assumptions about epochs and architecture design?”
LLM System Design #29
Apr 19, 2026Senior Computer Vision Engineer interview at Meta
The Generative Routing Trap
“Our e-commerce app needs an image translation feature to convert clothing images across 10 different seasonal and regional styles without paired data. How do you architect the generative routing?”
Deep Learning #24
Apr 14, 2026Senior AI Engineer interview at Meta
The Backprop Routing Trap
“You wrote a custom, bare-metal CUDA Max Pooling operation that cuts inference latency by 40%. But when you drop it into the training loop, gradient descent completely breaks. Why?”
Deep Learning #20
Apr 10, 2026Senior Computer Vision Engineer interview at Meta
The 1x1 Convolution Trap
“Your production CNN is hitting severe memory limits on your 80GB A100s. A junior engineer suggests replacing several 3x3 convolutions with 1x1 convolutions to “save space.”
Deep Learning #19
Apr 9, 2026Senior ML Engineer interview at Meta
The Convexity Assumption Trap
“You’re migrating a legacy continuous prediction model into a multi-class classifier. A junior dev suggests keeping the L2 (MSE) loss for the new Softmax outputs because ‘error is error.’ Why is this guaranteed to break the optimizer in production?”
Deep Learning #15
Apr 5, 2026Senior ML Engineer interview at Meta
The Dropout Scaling Trap
“You trained a large network with a heavy Dropout rate of 0.5. It performs flawlessly on the validation set. But when you export the raw weights to a custom offline C++ inference engine, the activations completely blow up and saturate. Assuming zero code bugs, what mathematical correction was missed?”
Deep Learning #14
Apr 4, 2026Senior Computer Vision Engineer interview at Meta
The Max Pooling Gradient Trap
“You’re using a Max activation function across a set of feature maps. During backpropagation debugging, you notice that the vast majority of your weights in the preceding layer aren’t updating at all. Why is this mathematically expected, and how does the engine handle exact ties?”
Deep Learning #10
Mar 31, 2026Senior ML Engineer interview at Meta
The Global Accuracy Trap
“You just bumped a model’s accuracy from 75% to 85%, crossing the business cutoff for deployment. In what scenario does deploying this mathematically ‘better’ model actually destroy the end-user experience?”
Deep Learning #5
Mar 26, 2026Senior ML Engineer interview at Meta
The I/O Starvation Trap
“You just migrated your team’s deep learning workloads from local hardware to a massive AWS GPU cluster to accelerate training. The expensive instances are successfully spinning, but your training iteration speed has actually flatlined. What is the hidden system bottleneck throttling your pipeline?”
Deep Learning #4
Mar 25, 2026Senior ML Engineer interview at Meta
The Leaderboard Overfitting Trap
“Your team just ensembled 12 different deep learning models to squeeze out an extra 2% accuracy and secure the top spot on our internal leaderboard. Why is directly deploying this ‘winning’ submission a terrible idea for our live system, and what technique do you use instead?”
Deep Learning #3
Mar 24, 2026Senior ML Engineer interview at Meta
The Memory Fragmentation Trap
“A junior dev hands you a 500-line PyTorch Out-of-Memory (OOM) stack trace and asks for help. What is your exact debugging workflow before you even think about telling them to ‘just lower the batch size’?”
Deep Learning #2
Mar 23, 2026Senior AI Engineer interview at Meta
The VRAM Bottleneck Trap
“Instead of relying on 𝘗𝘺𝘛𝘰𝘳𝘤𝘩'𝘴 𝘣𝘶𝘪𝘭𝘵-𝘪𝘯 𝘢𝘶𝘵𝘰𝘨𝘳𝘢𝘥 𝘦𝘯𝘨𝘪𝘯𝘦, in what highly constrained production scenario does writing 𝘤𝘶𝘴𝘵𝘰𝘮 𝘧𝘰𝘳𝘸𝘢𝘳𝘥 𝘢𝘯𝘥 𝘣𝘢𝘤𝘬𝘸𝘢𝘳𝘥 𝘱𝘢𝘴𝘴𝘦𝘴 𝘧𝘳𝘰𝘮 𝘴𝘤𝘳𝘢𝘵𝘤𝘩 become an absolute engineering necessity?”
Deep Learning #1
Mar 22, 2026Senior AI Engineer interview at Meta
The Multimodal Geometry Trap
“We are building a Multimodal LLM like LLaVA. We need to feed the frozen CLIP image embeddings into our Language Model. Should we use the final [CLS] token?”
Computer Vision #15
Jan 16, 2026Computer Vision Engineer interview at Meta
The Early vs Slow Fusion Trap
“We’re debating between 𝘌𝘢𝘳𝘭𝘺 𝘍𝘶𝘴𝘪𝘰𝘯 and 𝘚𝘭𝘰𝘸 𝘍𝘶𝘴𝘪𝘰𝘯 for our new video understanding model. Everyone knows 𝘚𝘭𝘰𝘸 𝘍𝘶𝘴𝘪𝘰𝘯 captures motion better, but what is the specific computational consequence of maintaining that temporal dimension through multiple layers that kills our training budget?”
Computer Vision #10
Jan 11, 2026Senior AI Engineer interview at Meta
The Quantization Gradient Trap
“We need to switch to 𝐐𝐮𝐚𝐧𝐭𝐢𝐳𝐚𝐭𝐢𝐨𝐧 𝐀𝐰𝐚𝐫𝐞 𝐓𝐫𝐚𝐢𝐧𝐢𝐧𝐠 (𝐐𝐀𝐓) because post-training quantization is tanking our accuracy. But the rounding operation (Float -> Int8) is a step function with a derivative of zero. How do you actually backpropagate gradients through it to update the weights?”
NLP #20
Dec 26, 2025Senior AI Engineer interview at Meta
The Positional Encoding Wall
“We used 𝘙𝘰𝘗𝘌 (𝘙𝘰𝘵𝘢𝘳𝘺 𝘗𝘰𝘴𝘪𝘵𝘪𝘰𝘯𝘢𝘭 𝘌𝘮𝘣𝘦𝘥𝘥𝘪𝘯𝘨𝘴) for Llama instead of standard absolute learned embeddings. Apart from the math, what is the critical advantage RoPE offers when we need to run inference on sequences longer than what we trained on?”
NLP #15
Dec 21, 2025AI Engineer interview at Meta
The Tokenization Brittleness Trap
“We deployed a Llama-3 based app. We removed a single whitespace in the prompt template, and our benchmark accuracy tanked by 12%. Why is the model so brittle to a simple format change, and why didn’t instruction tuning prevent this?”
NLP #14
Dec 20, 2025Senior ML Engineer interview at Meta
The Exploding Gradient Trap
“You’re training a 7B parameter Llama-style model. In the first 1000 steps, your gradients start oscillating wildly and the loss spikes. How do you fix it?”
NLP #7
Dec 13, 2025Senior ML System Desgin Engineer interview at Meta
The Silent Feature Death
“Your new pricing model has 99.9% availability, 15ms latency, and zero exceptions. Yet, the A/B test results are complete garbage. What happened?”
ML System Design #21
Dec 6, 2025Senior ML System Engineer interview at Meta
The LoRA Knowledge Trap
“We need to adapt Llama-3 70B to the highly technical Medical domain. We are GPU-constrained, so we can’t do full fine-tuning. How do we proceed?”
ML System Design #12
Nov 29, 2025Senior AI Engineer interview at Meta
The FLOPs Fallacy
“You’re A/B testing two 70B models - one Multi-Head Attention (MHA), one Grouped Query Attention (GQA). Your colleague argues they’ll have the same inference speed since FLOPs and parameter counts are identical. Is this assumption correct?”
LLM System Design #15
Nov 12, 2025AI Engineer interview at Meta
When Memory Becomes the Enemy
“We all know KV Caching speeds up token generation. What’s the primary bottleneck this technique creates in a high-throughput production system, and how do you conceptually solve it?”
LLM System Design #6
Nov 5, 2025