diff --git a/_docs/02_tasks/_dependencies_table.md b/_docs/02_tasks/_dependencies_table.md
index 7d05446..11cd88c 100644
--- a/_docs/02_tasks/_dependencies_table.md
+++ b/_docs/02_tasks/_dependencies_table.md
@@ -1,56 +1,44 @@
 # Dependencies Table
 
-**Date**: 2026-03-22
-**Total Tasks**: 11
-**Total Complexity Points**: 35
+**Date**: 2026-03-31
+**Total Tasks**: 15 (11 done + 4 new)
+**Total Complexity Points**: 43
 
-| Task | Name | Complexity | Dependencies | Epic |
-|------|------|-----------|-------------|------|
-| AZ-138 | test_infrastructure | 5 | None | AZ-137 |
-| AZ-139 | test_health_engine | 3 | AZ-138 | AZ-137 |
-| AZ-140 | test_single_image | 3 | AZ-138 | AZ-137 |
-| AZ-141 | test_tiling | 3 | AZ-138 | AZ-137 |
-| AZ-142 | test_async_sse | 3 | AZ-138 | AZ-137 |
-| AZ-143 | test_video | 3 | AZ-138, AZ-142 | AZ-137 |
-| AZ-144 | test_negative | 2 | AZ-138 | AZ-137 |
-| AZ-145 | test_resilience | 5 | AZ-138, AZ-142 | AZ-137 |
-| AZ-146 | test_performance | 3 | AZ-138 | AZ-137 |
-| AZ-147 | test_security | 2 | AZ-138 | AZ-137 |
-| AZ-148 | test_resource_limits | 3 | AZ-138, AZ-142 | AZ-137 |
+## Completed Tasks (Steps 4-6)
 
-## Test Scenario Coverage
+| Task | Name | Complexity | Dependencies | Epic | Status |
+|------|------|-----------|-------------|------|--------|
+| AZ-138 | test_infrastructure | 5 | None | AZ-137 | done |
+| AZ-139 | test_health_engine | 3 | AZ-138 | AZ-137 | done |
+| AZ-140 | test_single_image | 3 | AZ-138 | AZ-137 | done |
+| AZ-141 | test_tiling | 3 | AZ-138 | AZ-137 | done |
+| AZ-142 | test_async_sse | 3 | AZ-138 | AZ-137 | done |
+| AZ-143 | test_video | 3 | AZ-138, AZ-142 | AZ-137 | done |
+| AZ-144 | test_negative | 2 | AZ-138 | AZ-137 | done |
+| AZ-145 | test_resilience | 5 | AZ-138, AZ-142 | AZ-137 | done |
+| AZ-146 | test_performance | 3 | AZ-138 | AZ-137 | done |
+| AZ-147 | test_security | 2 | AZ-138 | AZ-137 | done |
+| AZ-148 | test_resource_limits | 3 | AZ-138, AZ-142 | AZ-137 | done |
 
-| Task | Scenarios |
-|------|-----------|
-| AZ-138 | Infrastructure scaffold (no test scenarios) |
-| AZ-139 | FT-P-01, FT-P-02, FT-P-14, FT-P-15 |
-| AZ-140 | FT-P-03, FT-P-05, FT-P-06, FT-P-07, FT-P-13 |
-| AZ-141 | FT-P-04, FT-P-16 |
-| AZ-142 | FT-P-08, FT-P-09, FT-N-04 |
-| AZ-143 | FT-P-10, FT-P-11, FT-P-12 |
-| AZ-144 | FT-N-01, FT-N-02, FT-N-03, FT-N-05 |
-| AZ-145 | FT-N-06, FT-N-07, NFT-RES-01, NFT-RES-02, NFT-RES-03, NFT-RES-04 |
-| AZ-146 | NFT-PERF-01, NFT-PERF-02, NFT-PERF-03, NFT-PERF-04 |
-| AZ-147 | NFT-SEC-01, NFT-SEC-02, NFT-SEC-03 |
-| AZ-148 | FT-N-08, NFT-RES-LIM-01, NFT-RES-LIM-02, NFT-RES-LIM-03, NFT-RES-LIM-04 |
+## New Tasks (Step 8 — Distributed Architecture)
 
-**Total scenarios**: 39/39 covered
+| Task | Name | Complexity | Dependencies | Epic | Status |
+|------|------|-----------|-------------|------|--------|
+| AZ-173 | stream_based_run_detect | 3 | None | AZ-172 | todo |
+| AZ-174 | db_driven_ai_config | 2 | None | AZ-172 | todo |
+| AZ-175 | media_table_integration | 2 | AZ-173 | AZ-172 | todo |
+| AZ-176 | cleanup_obsolete_path_code | 1 | AZ-173, AZ-174 | AZ-172 | todo |
 
-## Execution Order
+## Execution Order (New Tasks)
 
-**Batch 1** (no dependencies beyond infrastructure):
-- AZ-138: test_infrastructure (5 pts)
+**Batch 1** (no internal dependencies):
+- AZ-173: stream_based_run_detect (3 pts) — PyAV byte-stream decoding
+- AZ-174: db_driven_ai_config (2 pts) — fetch config from annotations service
 
-**Batch 2** (depends on AZ-138 only):
-- AZ-139: test_health_engine (3 pts)
-- AZ-140: test_single_image (3 pts)
-- AZ-141: test_tiling (3 pts)
-- AZ-142: test_async_sse (3 pts)
-- AZ-144: test_negative (2 pts)
-- AZ-146: test_performance (3 pts)
-- AZ-147: test_security (2 pts)
+**Batch 2** (depends on AZ-173):
+- AZ-175: media_table_integration (2 pts) — Media record lifecycle
 
-**Batch 3** (depends on AZ-138 + AZ-142):
-- AZ-143: test_video (3 pts)
-- AZ-145: test_resilience (5 pts)
-- AZ-148: test_resource_limits (3 pts)
+**Batch 3** (depends on AZ-173 + AZ-174):
+- AZ-176: cleanup_obsolete_path_code (1 pt) — remove old path-based code
+
+**New tasks total**: 4 tasks, 8 complexity points, 3 batches
diff --git a/_docs/02_tasks/backlog/AZ-173_stream_based_run_detect.md b/_docs/02_tasks/todo/AZ-173_stream_based_run_detect.md
similarity index 71%
rename from _docs/02_tasks/backlog/AZ-173_stream_based_run_detect.md
rename to _docs/02_tasks/todo/AZ-173_stream_based_run_detect.md
index 6c15155..30a9d3b 100644
--- a/_docs/02_tasks/backlog/AZ-173_stream_based_run_detect.md
+++ b/_docs/02_tasks/todo/AZ-173_stream_based_run_detect.md
@@ -38,16 +38,19 @@ Remove:
 - `paths` iteration loop in `run_detect`
 - Direct `cv2.VideoCapture(local_path)` and `cv2.imread(local_path)` calls
 
-## Video Stream Processing Options
+## Video Stream Processing — PyAV (preferred)
 
-**Option A: Write-then-read**
-Write entire upload to temp file, then open with `cv2.VideoCapture`. Simple but not real-time.
+Use **PyAV** (`av` package) for byte-stream video decoding. Preferred approach per user decision.
 
-**Option B: Concurrent pipe**
-One thread writes incoming bytes to a file; another thread reads frames via `cv2.VideoCapture` on the growing file. Requires careful synchronization.
+**Why PyAV**: Decode frames directly from bytes/stream via `av.open(io.BytesIO(data))` without intermediate file I/O. Most flexible for streaming, no concurrent file read/write synchronization needed.
 
-**Option C: PyAV byte-stream decoding**
-Use `av.open(io.BytesIO(data))` or a custom `av.InputContainer` to decode frames directly from bytes without file I/O. Most flexible for streaming.
+**Investigation needed during implementation**:
+- Verify PyAV can decode frames from a growing `BytesIO` or chunked stream (not just complete files)
+- Check if `av.open()` supports reading from an async generator or pipe for true streaming
+- Evaluate whether simultaneous disk write can be done with a `tee`-style approach (write bytes to file while feeding to PyAV)
+- Confirm PyAV frame format (RGB/BGR/YUV) compatibility with existing `process_frames` pipeline (expects BGR numpy arrays)
+- Check PyAV version compatibility with the project's Python/Cython setup
+- Fallback: if PyAV streaming is not viable, use write-then-read with `cv2.VideoCapture` on completed file
 
 ## Acceptance Criteria
 
diff --git a/_docs/02_tasks/backlog/AZ-174_db_driven_ai_config.md b/_docs/02_tasks/todo/AZ-174_db_driven_ai_config.md
similarity index 100%
rename from _docs/02_tasks/backlog/AZ-174_db_driven_ai_config.md
rename to _docs/02_tasks/todo/AZ-174_db_driven_ai_config.md
diff --git a/_docs/02_tasks/backlog/AZ-175_media_table_integration.md b/_docs/02_tasks/todo/AZ-175_media_table_integration.md
similarity index 100%
rename from _docs/02_tasks/backlog/AZ-175_media_table_integration.md
rename to _docs/02_tasks/todo/AZ-175_media_table_integration.md
diff --git a/_docs/02_tasks/backlog/AZ-176_cleanup_obsolete_path_code.md b/_docs/02_tasks/todo/AZ-176_cleanup_obsolete_path_code.md
similarity index 100%
rename from _docs/02_tasks/backlog/AZ-176_cleanup_obsolete_path_code.md
rename to _docs/02_tasks/todo/AZ-176_cleanup_obsolete_path_code.md
diff --git a/_docs/_autopilot_state.md b/_docs/_autopilot_state.md
index 4ec616d..d3cb6c4 100644
--- a/_docs/_autopilot_state.md
+++ b/_docs/_autopilot_state.md
@@ -2,53 +2,8 @@
 
 ## Current Step
 flow: existing-code
-step: 7
-name: Refactor
-status: completed
-sub_step: done
+step: 9
+name: Implement
+status: not_started
+sub_step: 0
 retry_count: 0
-
-## Completed Steps
-
-| Step | Name | Completed | Key Outcome |
-|------|------|-----------|-------------|
-| 1 | Document | 2026-03-21 | 10 modules, 4 components, full _docs/ generated from existing codebase |
-| 2 | Test Spec | 2026-03-21 | 39 test scenarios (16 positive, 8 negative, 11 non-functional), 85% total coverage, 5 artifacts produced |
-| 3 | Code Testability Rev. | 2026-03-29 | Engine factory refactoring completed: polymorphic EngineClass pattern (TensorRT/CoreML/ONNX) with auto-detection. Hardcoded values aligned with Docker compose. |
-| 4 | Decompose Tests | 2026-03-23 | 11 tasks (AZ-138..AZ-148), 35 complexity points, 3 batches. Phase 3 test data gate PASSED: 39/39 scenarios validated, 12 data files provided. |
-| 5 | Implement Tests | 2026-03-23 | 11 tasks implemented across 4 batches, 38 tests (2 skipped), all code reviews PASS_WITH_WARNINGS. Commits: 5418bd7, a469579, 861d4f0, f0e3737. |
-| 6 | Run Tests | 2026-03-30 | 23 passed, 0 failed, 0 skipped, 0 errors in 11.93s. Fixed: Cython __reduce_cython__ (clean rebuild), missing Pillow dep, relative MEDIA_DIR paths. Removed 14 dead/unreachable tests. Updated test-run skill to treat skips as blocking gate. |
-| 7 | Refactor | 2026-03-31 | Engine-centric dynamic batch refactoring. Moved source to src/. Engine pipeline redesign: preprocess/postprocess/process_frames in base InferenceEngine, dynamic batching per engine (CoreML=1, TensorRT=GPU-calculated, ONNX=config). Fixed: video partial batch flush, image accumulation data loss, frame-is-None crash. Removed detect_single_image (POST /detect delegates to run_detect). Dead code: removed msgpack, serialize methods, unused constants/fields. Configurable classes.json + log paths, HTTP timeouts. 28 e2e tests pass. |
-
-## Key Decisions
-- User chose to document existing codebase before proceeding
-- Component breakdown: 4 components (Domain, Inference Engines, Inference Pipeline, API)
-- Verification: 4 legacy issues found and documented (unused serialize/from_msgpack, orphaned queue declarations)
-- Input data coverage approved at ~90% (Phase 1a)
-- Test coverage approved at 85% (21/22 AC, 13/18 restrictions) with all gaps justified
-- User chose refactor path (decompose tests → implement tests → refactor)
-- Integration Tests Epic: AZ-137
-- Test Infrastructure: AZ-138 (5 pts)
-- 10 integration test tasks decomposed: AZ-139 through AZ-148 (30 pts)
-- Total: 11 tasks, 35 complexity points, 3 batches
-- Phase 3 (Test Data Validation Gate) PASSED: 39/39 scenarios have data, 85% coverage, 0 tests removed
-- Test data: 6 images, 3 videos, 1 ONNX model, 1 classes.json provided by user
-- User confirmed dependency table and test data gate
-- Jira MCP auth skipped — tickets not transitioned to In Testing
-- Test run: removed 14 dead/unreachable tests (explicit @skip + runtime always-skip), added .c to .gitignore
-- User chose to refactor (option A) — clean up legacy dead code
-- User requested: move code to src/, thorough re-analysis, exhaustive refactoring list
-- Refactoring round: 01-code-cleanup, automatic mode, 15 changes identified
-- User feedback: analyze logical flow contradictions, not just static code. Updated refactor skill Phase 1 with logical flow analysis.
-- User chose: split scope — engine refactoring as Step 7, architecture shift (streaming, DB config, media storage, Jetson) as Step 8
-- User chose: remove detect_single_image, POST /detect delegates to run_detect
-- GPU memory fraction: 80% for inference, 20% buffer (Jetson 40% deferred to Step 8)
-
-## Last Session
-date: 2026-03-31
-ended_at: Step 7 complete — all 11 todos done, 28 e2e tests pass
-reason: Refactoring complete
-notes: Engine-centric dynamic batch refactoring implemented. Source moved to src/. InferenceEngine base class now owns preprocess/postprocess/process_frames with per-engine max_batch_size. CoreML overrides preprocess (direct PIL, no blob reversal) and postprocess. TensorRT calculates max_batch_size from GPU memory (80% fraction) with optimization profiles for dynamic batch. All logical flow bugs fixed (LF-01 through LF-09). Dead code removed (msgpack, serialize, unused constants). POST /detect unified through run_detect. Next: Step 8 (architecture shift — streaming media, DB-backed config, media storage, Jetson support).
-
-## Blockers
-- none