Expert FDE interview questions

Human 1-5 scores are biased toward short, simple code. Debias the labels, fuse test-pass rate without letting the agent game it, and pick PPO with a fused reward over a DPO ranking loss. Plus the hard cap on the action space so the agent can't delete the repo.

Reward only on the final answer and learning crawls; shape intermediate tool calls and the model learns to spam SQL. Use outcome-grounded credit assignment, GRPO over a group of trajectories, and fine-tune a pre-aligned model so you're not teaching tool syntax from scratch.

A fixed 1000-token budget forces the search choice. Step-level beam search guided by a process-reward model beats best-of-N and MCTS on accuracy per token for math, and you collect the PRM labels automatically with Monte-Carlo rollouts, no human step annotation.

A multi-hop question over a property graph is a query-generation problem, not a similarity-search one. The staff-level move is grounding generation in the live schema and closing an auto-healing loop on the DB's own error messages.

You cannot unit-test a stochastic two-agent pipeline with a golden output. You evaluate it the way you stress a distributed system: thousands of simulated runs, controlled fault injection, and a judge calibrated against humans.

The hard parts are not the four roles. They are the communication substrate, a termination condition that is not just a retry cap, and stopping the Coder from forgetting the ticket after five repair cycles.

Deep-research agents run unsupervised for twenty minutes and their report goes straight to an executive. The failure mode is not a crash, it is confident citations that do not support the claims. Most candidates design the retrieval and skip the verification pass, the run budget, and the freshness story.

The citation is real and semantically relevant, yet the advice belongs to different equipment. Resolve applicability before ranking and preserve the version decision through generation and evaluation.

The customer wants the agent to practice before it can write to production. Design the task world, reset semantics, tool boundary and verifier, then explain when its scores justify training or release.

The naive pipeline gives you 10 million 'summarize this paragraph' pairs and a model that can only summarize paragraphs. The hard part is diversity and quality control at 100-billion-document scale, not the generation call.

Five LLM agents over a hundred steps, and the bar is zero contradictions with a 500-page contract. The real test is the state and grounding layer, not the agent prompts.

No data leaves, you cannot pre-inspect the corpus, and your toolchain arrives on a single USB. Design adaptation and evaluation when you will never observe a single input or output.

Ten X-ray studies a second, a 2-second budget, and a missed cancer is unacceptable. The hard parts are recall-first thresholding, a monitor that watches the radiologist, and surviving a new scanner.

Constrained JSON/grammar decoding usually means a Python callback per token, which serializes the GPU. The staff answer keeps the DFA on-device and turns the mask into a single coalesced gather.

Reprefilling the whole history every turn makes a long chat O(n^2). The real design is a radix tree of KV blocks with copy-on-write branching, a weight-epoch tag, and an eviction policy that protects live turns.

Each customer has their own LoRA on a shared 8B. The S-LoRA-style answer keeps one base in memory, pages adapters in and out, and tags every KV block with an adapter ID so reuse never crosses tenants.

A 7B VLM at 25 FPS on an embedded GPU sounds impossible until you stop processing every frame fully. The answer is token pruning, frame-to-frame KV reuse, and a three-stage pipeline that hides latency.

Going 4K to 1M is not one trick. It is RoPE rescaling chosen by frequency band, a long-doc continued-pretrain mix, needle evals that actually catch failures, and KV math that forces ring attention.

64 H100s for 30 days is a real budget with a real token count. The answer sizes the run from Chinchilla, picks parallelism for this exact cluster, and has a concrete plan for loss spikes and silent data corruption.

A red-team model that learns over many episodes which attacks bypass your filters is only useful if it explores instead of farming one exploit, and only safe if every discovered attack becomes target training data.

Two ways to prove a cloud LLM run is the audited model on the customer's input: a zero-knowledge proof of the computation, or a TEE attestation. They differ by roughly an order of magnitude in overhead and entirely in trust assumptions.

SAEs can surface candidate 'deception' features in the residual stream, but a correlated feature is not a cause. The real work, and the honest answer, is causal validation and admitting what interpretability cannot yet prove.

Fine-tune a LoRA adapter on the user's typing on the phone's neural engine, add noise locally with DP-SGD, and aggregate only adapter deltas, so no individual's vocabulary is recoverable from the global update.