infra(argo): alpha-perception workflow + submission script

Argo WorkflowTemplate at infra/k8s/argo/alpha-perception-template.yaml
runs the stacked Mamba2 -> CfC -> heads PerceptionTrainer on a single
L40S in fr-par-2. Two-stage DAG:
  ensure-binary  (ci-compile-cpu pool, sccache-backed cargo build of
                  alpha_train example, SHA-keyed binary cache under
                  /data/bin/$SHORT_SHA/)
  train          (ci-training-l40s pool, runs the cached binary against
                  /data/futures-baseline/mbp10 with predecoded sidecar
                  cache at /feature-cache/predecoded, writes
                  alpha_train_summary.json to
                  /feature-cache/alpha-perception-runs/$SHA/)

Defaults mirror the validated synthetic-overfit smoke config:
  epochs=5, seq_len=32, mamba2_state_dim=16, lr_cfc=3e-3,
  lr_mamba2=1e-3, n_train_seqs=8000, n_val_seqs=1000, seed=0x4242

Submission script scripts/argo-alpha-perception.sh wraps argo submit
with the standard L40S/H100 cuda-compute-cap mapping. --watch
follows logs.

Workflow nodeSelector pinned to fr-par-2 (consistent with the cluster
topology constraint). ttlStrategy 1h after completion;
activeDeadlineSeconds 4h cap (well above expected ~30-90 min wall).

This is the cluster entrypoint for the stacked perception design.
Once it lands a summary on MinIO, the gate runner (alpha_gate, Task
17) can consume it.

Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>

scripts/argo-alpha-perception.sh

@@ -0,0 +1,110 @@
+#!/usr/bin/env bash
+# Submit the alpha-perception workflow.
+#
+# Trains the stacked Mamba2 -> CfC -> heads perception model on MBP-10
+# from the training-data PVC. Emits alpha_train_summary.json with
+# per-horizon validation AUC to the feature-cache PVC.
+#
+# Defaults match the validated synthetic-overfit smoke config; cluster
+# runs can override for sweep work.
+#
+# Usage:
+#   ./scripts/argo-alpha-perception.sh                   # current HEAD on ml-alpha-phase-a
+#   ./scripts/argo-alpha-perception.sh --branch main
+#   ./scripts/argo-alpha-perception.sh --epochs 1 --n-train-seqs 1000  # quick smoke
+#   ./scripts/argo-alpha-perception.sh --watch
+
+set -euo pipefail
+
+SHA=HEAD
+BRANCH=$(git symbolic-ref --short HEAD 2>/dev/null || echo "ml-alpha-phase-a")
+GPU_POOL=ci-training-l40s
+EPOCHS=5
+SEQ_LEN=32
+MAMBA2_STATE_DIM=16
+LR_CFC=3e-3
+LR_MAMBA2=1e-3
+N_TRAIN_SEQS=8000
+N_VAL_SEQS=1000
+SEED=0x4242
+WATCH=false
+
+usage() {
+  cat <<EOF
+Usage: $0 [OPTIONS]
+  --sha <commit>            Git SHA (default: HEAD on the current branch)
+  --branch <branch>         Git branch (default: $BRANCH)
+  --gpu-pool <pool>         GPU pool (default: $GPU_POOL)
+  --epochs <n>              Training epochs (default: $EPOCHS)
+  --seq-len <n>             Snapshots per sequence window (default: $SEQ_LEN)
+  --mamba2-state-dim <n>    Mamba2 SSM state dim (default: $MAMBA2_STATE_DIM)
+  --lr-cfc <f>              CfC learning rate (default: $LR_CFC)
+  --lr-mamba2 <f>           Mamba2 learning rate (default: $LR_MAMBA2)
+  --n-train-seqs <n>        Train sequences per epoch (default: $N_TRAIN_SEQS)
+  --n-val-seqs <n>          Val sequences per epoch (default: $N_VAL_SEQS)
+  --seed <n>                Random seed (default: $SEED)
+  --watch                   Follow logs via argo watch
+EOF
+}
+
+while [[ $# -gt 0 ]]; do
+  case "$1" in
+    --sha)               SHA="$2"; shift 2 ;;
+    --branch)            BRANCH="$2"; shift 2 ;;
+    --gpu-pool)          GPU_POOL="$2"; shift 2 ;;
+    --epochs)            EPOCHS="$2"; shift 2 ;;
+    --seq-len)           SEQ_LEN="$2"; shift 2 ;;
+    --mamba2-state-dim)  MAMBA2_STATE_DIM="$2"; shift 2 ;;
+    --lr-cfc)            LR_CFC="$2"; shift 2 ;;
+    --lr-mamba2)         LR_MAMBA2="$2"; shift 2 ;;
+    --n-train-seqs)      N_TRAIN_SEQS="$2"; shift 2 ;;
+    --n-val-seqs)        N_VAL_SEQS="$2"; shift 2 ;;
+    --seed)              SEED="$2"; shift 2 ;;
+    --watch)             WATCH=true; shift ;;
+    -h|--help)           usage; exit 0 ;;
+    *) echo "Unknown option: $1"; usage; exit 1 ;;
+  esac
+done
+
+case "$GPU_POOL" in
+  ci-training-l40s)
+    SM=89 ;;
+  ci-training-h100*|ci-training-h100x2|ci-training-h100-sxm)
+    SM=90 ;;
+  *)
+    echo "Unknown gpu-pool: $GPU_POOL"
+    exit 1 ;;
+esac
+
+echo "Submitting alpha-perception workflow..."
+echo "  branch:           $BRANCH"
+echo "  sha:              $SHA"
+echo "  gpu-pool:         $GPU_POOL (sm_$SM)"
+echo "  epochs:           $EPOCHS"
+echo "  seq-len:          $SEQ_LEN"
+echo "  mamba2-state-dim: $MAMBA2_STATE_DIM"
+echo "  lr-cfc:           $LR_CFC"
+echo "  lr-mamba2:        $LR_MAMBA2"
+echo "  n-train-seqs:     $N_TRAIN_SEQS"
+echo "  n-val-seqs:       $N_VAL_SEQS"
+echo "  seed:             $SEED"
+
+WATCH_FLAG=""
+if [[ "$WATCH" == "true" ]]; then
+  WATCH_FLAG="--watch"
+fi
+
+argo submit -n foxhunt --from=wftmpl/alpha-perception \
+  -p commit-sha="$SHA" \
+  -p git-branch="$BRANCH" \
+  -p cuda-compute-cap="$SM" \
+  -p gpu-pool="$GPU_POOL" \
+  -p epochs="$EPOCHS" \
+  -p seq-len="$SEQ_LEN" \
+  -p mamba2-state-dim="$MAMBA2_STATE_DIM" \
+  -p lr-cfc="$LR_CFC" \
+  -p lr-mamba2="$LR_MAMBA2" \
+  -p n-train-seqs="$N_TRAIN_SEQS" \
+  -p n-val-seqs="$N_VAL_SEQS" \
+  -p seed="$SEED" \
+  $WATCH_FLAG