diff --git a/.github/workflows/ci.yml b/.github/workflows/ci.yml
index 005307934..8fc9a05f1 100644
--- a/.github/workflows/ci.yml
+++ b/.github/workflows/ci.yml
@@ -84,10 +84,17 @@ jobs:
             --ignore-vuln CVE-2025-45768
 
   backend-tests:
-    name: Backend Tests
+    name: Backend Tests (shard ${{ matrix.shard }}/4)
     runs-on: ubuntu-latest
     if: github.event_name == 'push' || github.event.pull_request.user.login != github.repository_owner
     needs: backend-lint
+    strategy:
+      # Don't cancel sibling shards if one fails — we want every shard's
+      # failure list, not just the first one, so a single PR push shows
+      # all broken tests in one go.
+      fail-fast: false
+      matrix:
+        shard: [1, 2, 3, 4]
     steps:
       - uses: actions/checkout@v4
 
@@ -110,11 +117,22 @@ jobs:
           pip install -r requirements.txt
           pip install -r requirements-dev.txt
 
-      - name: Run tests
+      - name: Run tests (shard ${{ matrix.shard }}/4)
         timeout-minutes: 10
         run: |
           cd backend
-          python -m pytest tests/ -v --tb=short --timeout=60 --timeout-method=thread -n auto
+          # -v dropped: 5300+ "PASSED foo::bar" lines per worker eat 30-60s
+          # of stdout I/O time on 2-vCPU runners. --tb=short is enough.
+          # --splits 4 --group N uses pytest-split to slice the collected
+          # test set roughly evenly across the 4 matrix shards; first run
+          # is name-hash-based, subsequent runs improve via .test_durations
+          # if you ever commit one (we don't — even the naive hash split
+          # gets us ≈25% per shard given the test mix here).
+          python -m pytest tests/ \
+            --tb=short \
+            --timeout=60 --timeout-method=thread \
+            -n auto \
+            --splits 4 --group ${{ matrix.shard }}
 
   # ============================================================================
   # Frontend Checks
diff --git a/backend/=0.9.0 b/backend/=0.9.0
new file mode 100644
index 000000000..e69de29bb
diff --git a/backend/tests/unit/test_code_quality.py b/backend/tests/unit/test_code_quality.py
index 73734e618..b8b864201 100644
--- a/backend/tests/unit/test_code_quality.py
+++ b/backend/tests/unit/test_code_quality.py
@@ -218,9 +218,27 @@ class TestModuleImports:
         """Verify all Python modules can be imported without errors.
 
         This catches syntax errors and missing dependencies.
+
+        IMPORTANT: We must NOT ``del sys.modules[name]`` to force a fresh
+        import here. ``backend.app.main`` is a stateful module — re-importing
+        it builds NEW module-level dicts (_timelapse_baselines,
+        _expected_prints, _active_prints, …) and re-runs ``root_logger.
+        addHandler(console_handler)``. Any test that already bound those
+        names via ``from backend.app.main import _timelapse_baselines`` now
+        holds a stale reference, while production code resolves the symbol
+        through the new module instance — they're two different dicts. CI
+        under -n 2 puts test_code_quality.py on the same worker as
+        test_print_start_assigns_printer_id_to_vp_archive.py and
+        test_timelapse_baseline_restart_recovery.py, and those tests see
+        their mock_archive un-mutated / their baseline dict empty even
+        though production logged the mutations went through. Local -n 30
+        spreads the tests across workers and the collision never happens.
+
+        ``importlib.import_module`` already covers the "is this importable"
+        check — it returns the cached module if cached, or runs the import
+        machinery if not. Either way, an import-time error surfaces here.
         """
         import importlib
-        import sys
 
         # Modules to test importing
         modules = [
@@ -235,9 +253,6 @@ class TestModuleImports:
         errors = []
         for module_name in modules:
             try:
-                # Remove from cache first to ensure fresh import
-                if module_name in sys.modules:
-                    del sys.modules[module_name]
                 importlib.import_module(module_name)
             except Exception as e:
                 errors.append(f"{module_name}: {type(e).__name__}: {e}")
diff --git a/requirements-dev.txt b/requirements-dev.txt
index 21e36dc9a..5e97229ba 100644
--- a/requirements-dev.txt
+++ b/requirements-dev.txt
@@ -4,6 +4,10 @@ pytest-asyncio>=0.23.0
 pytest-cov>=4.1.0
 pytest-xdist>=3.5.0
 pytest-timeout>=2.4.0
+# Test-suite sharding for CI matrix. Splits the test set evenly across N
+# shards via --splits/--group; falls back to test-name hashing on the first
+# run, and uses recorded durations on subsequent runs (.test_durations).
+pytest-split>=0.9.0
 httpx>=0.27.0
 ruff>=0.8.0
 pre-commit>=4.0