feat: add ML-based adaptive timeout prediction using LinearRegressor

Train a linear regression model on actual message delivery times to
predict tighter timeouts, replacing worst-case physics estimates.
Features: path length, message bytes, seconds since last RX, flood mode.
Global model with per-contact blending after 10+ observations per contact.
Falls back to existing physics formula when model has insufficient data.

lib/models/delivery_observation.dart

@@ -0,0 +1,43 @@
+class DeliveryObservation {
+  final String contactKey;
+  final int pathLength;
+  final int messageBytes;
+  final int secondsSinceLastRx;
+  final bool isFlood;
+  final int deliveryMs;
+  final DateTime timestamp;
+
+  DeliveryObservation({
+    required this.contactKey,
+    required this.pathLength,
+    required this.messageBytes,
+    required this.secondsSinceLastRx,
+    required this.isFlood,
+    required this.deliveryMs,
+    required this.timestamp,
+  });
+
+  Map<String, dynamic> toJson() {
+    return {
+      'contact_key': contactKey,
+      'path_length': pathLength,
+      'message_bytes': messageBytes,
+      'seconds_since_last_rx': secondsSinceLastRx,
+      'is_flood': isFlood,
+      'delivery_ms': deliveryMs,
+      'timestamp': timestamp.toIso8601String(),
+    };
+  }
+
+  factory DeliveryObservation.fromJson(Map<String, dynamic> json) {
+    return DeliveryObservation(
+      contactKey: json['contact_key'] as String,
+      pathLength: json['path_length'] as int,
+      messageBytes: json['message_bytes'] as int,
+      secondsSinceLastRx: json['seconds_since_last_rx'] as int? ?? 0,
+      isFlood: json['is_flood'] as bool,
+      deliveryMs: json['delivery_ms'] as int,
+      timestamp: DateTime.parse(json['timestamp'] as String),
+    );
+  }
+}