diff --git a/net-guardia/src/core/ml/model_loader.rs b/net-guardia/src/core/ml/model_loader.rs
index 55d7133..394bae5 100644
--- a/net-guardia/src/core/ml/model_loader.rs
+++ b/net-guardia/src/core/ml/model_loader.rs
@@ -1,9 +1,13 @@
 use std::path::PathBuf;
+use std::time::Instant;
+
+use macros::log;
 use tract_onnx::prelude::*;
 
 use crate::infrastructure::app_config::AppConfig;
 use crate::model::detection::ml_detection::RunnableModel;
 use crate::model::error::ml::MLError;
+use crate::model::log::ml::MLLog;
 use crate::model::system::config::MLInferenceConfig;
 
 pub struct MLModels {
@@ -36,13 +40,26 @@ impl MLModels {
     fn loader(model: &str, features: usize, batch_size: usize) -> Result<RunnableModel, MLError> {
         let model_path = PathBuf::from("models").join(model);
 
+        log!(MLLog::ModelLoading(
+            model.to_string(),
+            features,
+            batch_size
+        ));
+
+        let start = Instant::now();
+
         let load = || -> Result<RunnableModel, Box<dyn std::error::Error>> {
             let mut model = onnx().model_for_path(&model_path)?;
             model.set_input_fact(0, f32::fact([batch_size, features]).into())?;
             Ok(model.into_optimized()?.into_runnable()?)
         };
 
-        load().map_err(|_| MLError::ModelLoadFailed(model_path))
+        let result = load().map_err(|_| MLError::ModelLoadFailed(model_path));
+
+        let elapsed_ms = start.elapsed().as_millis() as u64;
+        log!(MLLog::ModelLoadComplete(model.to_string(), elapsed_ms));
+
+        result
     }
 
     pub fn get_model_info(&self, name: &str) -> String {
diff --git a/net-guardia/src/model/log/ml.rs b/net-guardia/src/model/log/ml.rs
index 865a877..e8e5cd6 100644
--- a/net-guardia/src/model/log/ml.rs
+++ b/net-guardia/src/model/log/ml.rs
@@ -45,6 +45,12 @@ loggable! {
         #[error("ML circuit breaker RESET: inference re-enabled after {cooldown_secs}s cooldown")]
         CircuitBreakerReset { cooldown_secs: u64 } => tracing::Level::WARN,
 
+        #[error("Loading ONNX model '{name}' (features={features}, batch_size={batch_size})...")]
+        ModelLoading { name: String, features: usize, batch_size: usize } => tracing::Level::INFO,
+
+        #[error("Model '{name}' loaded and optimized in {elapsed_ms}ms")]
+        ModelLoadComplete { name: String, elapsed_ms: u64 } => tracing::Level::INFO,
+
         #[error("Model watcher started, monitoring models/ for .onnx changes")]
         ModelWatcherStarted => tracing::Level::INFO,