What is model documentation (model card, datasheet)?
Machine Learning / AI
When is MAB better than A/B, and when is it worse?
What is BM25 and what are the parameters (k1, b)?
What is data versioning and how does it compare DVC, LakeFS, lakeFS?
What practices are used when synchronizing training-serving features?
How does SDXL differ architecturally from SD 1.5?
What are cuBLAS and cuDNN?
What is a decision transformer and how does it translate RL into supervised learning?
What is FSDP and how does it differ from DDP?
What is mean reciprocal rank (MRR) and where is it used in search evaluation?
What are temporal knowledge graphs?
What features does anomaly detection in time series (STL residuals, ESD) have?
What is weight tying between embedding and output projection?
What are concept activation vectors (TCAV)?
How to test an LLM agent (deterministic tests, replay, mocked tools)?
What is long-context evaluation? What is needle-in-haystack?
What is sim-to-real transfer and domain randomization?
How does SARIMA differ from ARIMA?
What is NGBoost and probabilistic predictions?
What is InstructPix2Pix?