Approximately how many tokens were there in one end-to-end pass?
Middle+
276
Machine Learning / AI
3611
0
What tool was used for structural document chunking?
Middle+
210
What is class imbalance in NER and how to work with it?
Middle+
197
What are the disadvantages of a multi-agent approach?
Middle+
196
What model was used under the hood of the ReAct agent and why was it chosen?
Middle+
196
How many candidates did you send to the reranker and how many were selected afterwards?
Middle+
191
Was it a JSON model or JSON schema through the API? What is the difference in usage?
Middle+
185
How were the BM25 values and dense embeddings combined before passing to the reranker (hybridization)?
Middle+
184
What is the difference between a cross-encoder (ranker) and a bi-encoder?
Middle+
183
Tell me about hybrid retrieval in RAG — which model did you choose for dense vectors?
Middle+
180
What kind of chunking was used and what type of documents were they (Word, PDF, Excel, etc.)?
Middle+
177
Why did you choose Qdrant as a vector storage?
Middle+
174
Was the model within its own boundary or did it communicate via an external API?
Middle+
172
How do you validate function call arguments before execution?
Middle+
172
Describe end-to-end, briefly what each of the four agents of the multi-agent system does.
Middle+
170
Do you check the code generated by the AI tool yourself afterward?
Middle+
169
What metrics can be used to evaluate NER on documents?
Middle+
167
What do you mean by an error (in the prompt approach)?
Middle+
165
How did you work with queries in the vector storage (Qdrant) and their impact on the agent's latency?
Middle+
164
What was the object of the state (TypedDict or other)?
Middle+
161