diff --git a/.dvc/.gitignore b/.dvc/.gitignore
index 6434553..f3d524f 100644
--- a/.dvc/.gitignore
+++ b/.dvc/.gitignore
@@ -5,4 +5,5 @@
 /updater.lock
 /state-journal
 /state-wal
-/cache
\ No newline at end of file
+/cache
+/tmp
diff --git a/README.md b/README.md
index 6a66131..e4759ee 100644
--- a/README.md
+++ b/README.md
@@ -23,6 +23,11 @@ Instructions for each exercise can be found under the
 *WARNING: the exercises build on top of each other, so you will not be able to
 skip steps ahead without executing them.*
 
+*WARNING 2: the workshop requires infrastructure that we only provision when
+needed, therefore you won't be able to execute the exercises on your own that
+require that shared infrastructure. We are working on a setup that allows
+running the workshop locally, but that is work in progress.*
+
 ## The Machine Learning Problem
 
 We built a simplified solution to a Kaggle problem posted by Corporación Favorita,
diff --git a/requirements.txt b/requirements.txt
index fe0f84a..d0dbff7 100644
--- a/requirements.txt
+++ b/requirements.txt
@@ -1,12 +1,13 @@
-dvc[gs]==0.35.7
-numpy==1.16.3
-pandas==0.24.2
-pylint==2.1.1
-pytest==4.4.1
-scikit-learn==0.20.3
-flask==1.0.2
-mlflow==0.9.1
-lime==0.1.1.33
-fluent-logger==0.9.3
+dvc[gs]==0.70.0
+numpy==1.17.4
+pandas==0.25.3
+pylint==2.4.4
+pytest==5.3.0
+scikit-learn==0.21.3
+flask==1.1.1
+mlflow==1.4.0
+lime==0.1.1.36
+fluent-logger==0.9.4
 s3fs==0.1.2
 seaborn==0.9.0
+joblib==0.14.0
diff --git a/src/app.py b/src/app.py
index 2308c1f..2f6527c 100644
--- a/src/app.py
+++ b/src/app.py
@@ -1,6 +1,6 @@
 from flask import Flask, render_template, jsonify, request
 from datetime import datetime
-from sklearn.externals import joblib
+import joblib
 import pandas as pd
 import decision_tree
 import os
diff --git a/src/decision_tree.py b/src/decision_tree.py
index 45fc115..99ed89a 100644
--- a/src/decision_tree.py
+++ b/src/decision_tree.py
@@ -3,7 +3,7 @@
 import pandas as pd
 import sys, os, json
 from sklearn.preprocessing import LabelEncoder
-from sklearn.externals import joblib
+import joblib
 sys.path.append(os.path.join('..', 'src'))
 sys.path.append(os.path.join('src'))
 from sklearn import tree, ensemble, metrics