diff --git a/docker/Dockerfile b/docker/Dockerfile
index 71216b42..5eac172a 100644
--- a/docker/Dockerfile
+++ b/docker/Dockerfile
@@ -12,14 +12,14 @@ ENV UV_COMPILE_BYTECODE=1 \
 # Copy workspace member directories first (needed for workspace resolution)
 COPY pyproject.toml uv.lock ./
 COPY haiku_rag_slim/pyproject.toml haiku_rag_slim/README.md haiku_rag_slim/LICENSE haiku_rag_slim/
-COPY evaluations/pyproject.toml evaluations/
 RUN --mount=type=cache,target=/root/.cache/uv \
-    uv sync --frozen --no-install-project
+    uv sync --frozen --no-install-project --no-dev
 
 # Install the project itself
-COPY . .
+COPY haiku_rag_slim haiku_rag_slim/
+COPY README.md LICENSE ./
 RUN --mount=type=cache,target=/root/.cache/uv \
-    uv sync --frozen --no-editable
+    uv sync --frozen --no-editable --no-dev
 
 # Final layer
 FROM python:3.13-slim
diff --git a/evaluations/evaluations/benchmark.py b/evaluations/evaluations/benchmark.py
index 9f3f4eb4..f75bd2bf 100644
--- a/evaluations/evaluations/benchmark.py
+++ b/evaluations/evaluations/benchmark.py
@@ -5,6 +5,10 @@ from typing import Any, cast
 
 import logfire
 import typer
+from haiku.rag.client import HaikuRAG
+from haiku.rag.config import AppConfig, find_config_file, load_yaml_config
+from haiku.rag.logging import configure_cli_logging
+from haiku.rag.qa import get_qa_agent
 from pydantic_ai.models.openai import OpenAIChatModel
 from pydantic_ai.providers.ollama import OllamaProvider
 from pydantic_evals import Dataset as EvalDataset
@@ -17,10 +21,6 @@ from evaluations.config import DatasetSpec
 from evaluations.datasets import DATASETS
 from evaluations.llm_judge import ANSWER_EQUIVALENCE_RUBRIC
 from evaluations.prompts import WIX_SUPPORT_PROMPT
-from haiku.rag.client import HaikuRAG
-from haiku.rag.config import AppConfig, find_config_file, load_yaml_config
-from haiku.rag.logging import configure_cli_logging
-from haiku.rag.qa import get_qa_agent
 
 QA_JUDGE_MODEL = "qwen3"