fix(ci): update Argo DQN training workflow for f32 pipeline

- train-dqn: hyperopt-trials 30→50, hyperopt-epochs 8→20,
  train-epochs 80→200 (matches dqn-production.toml, 24D search space)
- training-workflow: fixed --max-steps-per-epoch bug (was passing
  train-epochs as max-steps, limiting training to 80 steps/epoch
  instead of 80 epochs). Changed to --epochs.
- Added FOXHUNT_TRAINING_PROFILE env var pointing to
  dqn-production.toml so the training binary uses the correct config
  (adam_epsilon=1e-8, cosine LR, drawdown penalty, etc.)

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

infra/k8s/argo/train-dqn-template.yaml

@@ -41,11 +41,11 @@ spec:
       - name: gpu-pool
         value: ci-training-h100
       - name: hyperopt-trials
-        value: "30"          # 28D param space needs more exploration
+        value: "50"          # 24D param space needs full exploration
       - name: hyperopt-epochs
-        value: "8"
+        value: "20"          # 20 epochs per trial with early stopping
       - name: train-epochs
-        value: "80"           # DQN benefits from longer training
+        value: "200"          # matches dqn-production.toml
       - name: symbol
         value: ES.FUT
       - name: data-dir

infra/k8s/argo/training-workflow-template.yaml

@@ -338,7 +338,7 @@ spec:
               --mbp10-data-dir {{workflow.parameters.mbp10-data-dir}} \
               --trades-data-dir {{workflow.parameters.trades-data-dir}} \
               --output-dir /workspace/output \
-              --max-steps-per-epoch {{workflow.parameters.train-epochs}} \
+              --epochs {{workflow.parameters.train-epochs}} \
               $HYPEROPT_FLAG \
               --ensemble-top-k {{workflow.parameters.ensemble-top-k}}
 
@@ -349,6 +349,8 @@ spec:
             value: info
           - name: SQLX_OFFLINE
             value: "true"
+          - name: FOXHUNT_TRAINING_PROFILE
+            value: /workspace/config/training/dqn-production.toml
           - name: OTEL_EXPORTER_OTLP_ENDPOINT
             value: "http://tempo.foxhunt.svc.cluster.local:4317"
           - name: CUBLAS_WORKSPACE_CONFIG