Databricks Certified Context Engineer Associate - Certified Context Engineer Associate Exam
Page: 2 / 10
Total 50 questions
Question #6 (Topic: Exam A)
A context engineer is reviewing MLflow experiment tracking results comparing two system prompt configurations for a Databricks agent built to assist with contract review. Configuration P uses a role declaration, a full task instruction, and a dedicated edge-case handling section that describes in prose how the agent should respond to ambiguous or overlapping clause types, averaging 205 tokens per call and an 81% clause-classification accuracy. Configuration Q retains the role declaration, removes the prose edge-case handling section, shortens the task instruction to a concise directive, and adds three few-shot examples each demonstrating how the agent should classify a distinct edge-case clause type, averaging 340 tokens per call and a 93% clause-classification accuracy. The token budget increases by 135 tokens per call between Configuration P and Configuration Q. Latency increases from 290 ms to 380 ms.
Which conclusion is supported by these experiment tracking results?
Which conclusion is supported by these experiment tracking results?
A. The token increase is justified by the 12-percentage-point accuracy gain, and the shortened task instruction is the specific prompt element driving the performance improvement.
B. The token increase is justified by the 12-percentage-point accuracy gain, and the few-shot edge-case examples are the specific prompt element driving the performance improvement.
C. The token increase is not justified because removing the prose edge-case handling section and replacing it with examples introduces ambiguity that accuracy gain does not sufficiently offset.
D. The token increase is justified by the accuracy gain, but the improvement is attributable to the larger token budget because the removed edge-case handling section and added few-shot examples represent an equivalent token exchange.
Answer: B
Question #7 (Topic: Exam A)
A context engineer operates a finance Q&A agent on Databricks. The agent invokes a Genie space backed by tables in the finance.silver schema, and Genie selects which table to query for each user question by reading Unity Catalog table and column metadata. Two tables exist in the schema: revenue and revenue_gross. MLflow evaluation runs show the Correctness scorer flagging frequent failures where Genie queries revenue_gross when the user asks about net revenue, and queries revenue when the user asks about gross figures. Both tables are in active production use and must remain available to the Genie space, and the same disambiguation should be usable by other Genie spaces and BI consumers building against the same schema.
Which intervention resolves the failure?
Which intervention resolves the failure?
A. Add COMMENT text on both tables and their columns in Unity Catalog that clearly distinguishes the two tables – what each represents, the units, and the basis, so Genie reads the disambiguating descriptions when routing each user question.
B. Apply a Unity Catalog certified-status tag to revenue and a deprecated-status tag to revenue_gross so that downstream consumers treat revenue as the authoritative source.
C. Add a Genie space instruction at the space level stating which table to use for net-revenue questions and which to use for gross-revenue questions, so Genie applies the rule when routing each user question.
D. Apply a custom Unity Catalog tag categorizing one table as the primary revenue source and the other as a secondary view, so Genie’s tool selection prefers the primary.
Answer: A
Question #8 (Topic: Exam A)
A context engineer is building a Databricks agent for a financial research platform that answers analyst queries using a vector search index over earnings reports. The agent uses a 180-token system prompt, a 250-token conversation history window, and a RAG pipeline that retrieves 8 document chunks at runtime. Each chunk is 550 tokens, contributing 4,400 tokens of retrieved context. The model’s context window is 8,192 tokens, leaving only 3,362 tokens for the analyst query and model response. Profiling reveals that retrieved chunks consistently contain repetitive regulatory boilerplate and duplicated table headers that are rarely cited in the model’s final answers.
Which two changes should the context engineer make to reduce the token footprint of retrieved context without discarding relevant financial content? (Choose two.)
Which two changes should the context engineer make to reduce the token footprint of retrieved context without discarding relevant financial content? (Choose two.)
A. Reduce retrieved chunks from 8 to 3 and apply a reranker to select the highest-relevance chunks, preserving pertinent financial content while cutting injected tokens by over 60%.
B. Switch the retrieval strategy from cosine similarity to MMR to increase diversity across retrieved chunks, reducing redundancy between chunks without modifying chunk size or removing boilerplate.
C. Reduce each chunk from 550 tokens to 150-200 tokens by splitting at semantic boundaries and stripping repeated regulatory boilerplate and duplicate headers before indexing.
D. Re-index the earnings reports using a finer-grained embedding model to improve retrieval precision, surfacing more relevant passages without altering chunk boundaries or stripping metadata.
E. Expand the conversation history window from 250 tokens to 900 tokens to improve multi-turn coherence across analyst sessions and offset token pressure from the retrieved chunks.
Answer: AC
Question #9 (Topic: Exam A)
A context engineer is evaluating a Databricks agent deployed for a manufacturing company’s field service team. The agent uses a DELTA_SYNC Databricks AI Search index to search a knowledge base of equipment troubleshooting guides. The product line includes closely related models (CX-400, CX-410, CX-420) whose guides share similar procedural language but specify different part numbers, torque values, and calibration sequences for each model. The index uses databricks-gte-large-en embeddings with a chunk_size of 512 tokens and chunk_overlap of 50 tokens. Evaluation reveals that when a technician queries for a specific model, such as “CX-400 belt tensioner replacement procedure,” the agent frequently returns procedures from a different model in the same product family. The returned content appears relevant but contains incorrect part numbers and specifications for the queried model.
Which Databricks AI Search configuration change directly addresses this retrieval pattern?
Which Databricks AI Search configuration change directly addresses this retrieval pattern?
A. Set query_type to HYBRID so that the search combines semantic matching with keyword matching, merging both result sets with Reciprocal Rank Fusion.
B. Switch the embedding model to databricks-bge-large-en and configure the recommended query instruction to improve alignment between technician queries and guide passages.
C. Reduce chunk_size from 512 to 128 tokens so that each chunk contains a narrower passage where model-specific details like part numbers dominate the embedding.
D. Add columns_to_rerank with the built-in reranker to reorder initial search results so that the most model-specific passages rank highest.
Answer: A
Question #10 (Topic: Exam A)
A data analyst is building a Genie space for a hospital operations team that needs to analyze emergency department wait times, patient discharge rates, and nurse-to-patient ratios. The engineer has added the relevant tables, written general instructions describing the hospital operations domain, and configured sample questions. Stakeholders require that any metric involving nurse-to-patient ratios must use the official staffing calculation defined by the compliance team, but Genie should still answer free-form questions about discharge rates and wait times using ad hoc SQL when no trusted asset covers the question.
Which approach should the engineer take to satisfy both requirements?
Which approach should the engineer take to satisfy both requirements?
A. Add the compliance-approved nurse-to-patient ratio calculation to the general instructions as a text description so Genie applies the correct formula when generating ad hoc SQL for staffing questions.
B. Register example SQL queries as trusted assets for nurse-to-patient ratios, discharge rates, and wait time metrics so that every operational response earns the “Trusted” label and stakeholders can rely on all outputs.
C. Add the nurse-to-patient ratio query as a sample question paired with the expected answer so analysts see the correct format and Genie uses the sample as a template when generating ratio queries.
D. Register an example SQL query containing the compliance-approved nurse-to-patient ratio calculation as a trusted asset, and leave discharge rate and wait time queries unregistered so Genie generates ad hoc SQL for those questions.
Answer: D
