Add multilingual language and systems learning tracks

en/expanded-25-1/README.md

@@ -0,0 +1,9 @@
+# AI Runtime Engineering · 1
+
+[Agila University track](https://edu.agilaos.com/track.php?track=25&lang=en)
+
+About 15 minutes per lesson. Predict, run, change, explain, then complete the check in University.
+
+- [Lesson 1: What an AI runtime does](lesson-1/lesson.html)
+- [Lesson 2: Shapes before matrix multiplication](lesson-2/lesson.html)
+- [Lesson 3: Numerical tolerance and a reference](lesson-3/lesson.html)

en/expanded-25-1/lesson-1/lesson.html

@@ -0,0 +1,6 @@
+<!doctype html><html lang="en"><meta charset="utf-8"><title>What an AI runtime does</title><h1>What an AI runtime does</h1><p>15-minute session: read for 3 minutes, trace for 3, experiment for 7, and complete the check for 2. Spend longer whenever you need.</p><p>A model describes parameters and operations. A runtime loads that representation, validates shapes and types, allocates memory, schedules operations, and executes kernels. Training and inference are different workflows. Begin with a tiny deterministic computation so correctness can be checked by hand before introducing acceleration.</p><p><strong>Practice and evidence:</strong> Predict the example or draw its state transitions. Change one assumption, test the boundary, and record the actual result. Explain what this small model does not establish about a production system.</p><pre><code class="language-python">x=[2.0,3.0]
+w=[0.5,2.0]
+bias=1.0
+y=sum(a*b for a,b in zip(x,w))+bias
+print(y)
+</code></pre><p><strong>Expected result:</strong> <code>8.0</code></p><h3>Practice and evidence</h3><p>Keep three notes: your prediction, the actual result, and why they match or differ. The completion check grades understanding; the practice task builds skill.</p><p><a href="https://git.agilaos.com/matt/open_agx">Language reference: git.agilaos.com</a></p><p><a href="https://onnx.ai/onnx/intro/">Language reference: onnx.ai</a></p></html>

en/expanded-25-1/lesson-1/main.py

@@ -0,0 +1,5 @@
+x=[2.0,3.0]
+w=[0.5,2.0]
+bias=1.0
+y=sum(a*b for a,b in zip(x,w))+bias
+print(y)

en/expanded-25-1/lesson-2/lesson.html

@@ -0,0 +1,6 @@
+<!doctype html><html lang="en"><meta charset="utf-8"><title>Shapes before matrix multiplication</title><h1>Shapes before matrix multiplication</h1><p>15-minute session: read for 3 minutes, trace for 3, experiment for 7, and complete the check for 2. Spend longer whenever you need.</p><p>For A with shape m×k and B with shape k×n, the result has shape m×n. The inner dimensions must match. Check empty and ragged inputs before indexing. A shape check prevents some errors but cannot establish that the weights mean the right thing. The example is a small CPU reference, not a trained language model.</p><p><strong>Practice and evidence:</strong> Predict the example or draw its state transitions. Change one assumption, test the boundary, and record the actual result. Explain what this small model does not establish about a production system.</p><pre><code class="language-python">a=[[1,2],[3,4]]
+b=[[5],[6]]
+assert all(len(row)==len(b) for row in a)
+out=[[sum(a[i][k]*b[k][j] for k in range(len(b))) for j in range(len(b[0]))] for i in range(len(a))]
+print(out)
+</code></pre><p><strong>Expected result:</strong> <code>[[17], [39]]</code></p><h3>Practice and evidence</h3><p>Keep three notes: your prediction, the actual result, and why they match or differ. The completion check grades understanding; the practice task builds skill.</p><p><a href="https://git.agilaos.com/matt/open_agx">Language reference: git.agilaos.com</a></p><p><a href="https://onnx.ai/onnx/intro/">Language reference: onnx.ai</a></p></html>

en/expanded-25-1/lesson-2/main.py

@@ -0,0 +1,5 @@
+a=[[1,2],[3,4]]
+b=[[5],[6]]
+assert all(len(row)==len(b) for row in a)
+out=[[sum(a[i][k]*b[k][j] for k in range(len(b))) for j in range(len(b[0]))] for i in range(len(a))]
+print(out)

en/expanded-25-1/lesson-3/lesson.html

@@ -0,0 +1,4 @@
+<!doctype html><html lang="en"><meta charset="utf-8"><title>Numerical tolerance and a reference</title><h1>Numerical tolerance and a reference</h1><p>15-minute session: read for 3 minutes, trace for 3, experiment for 7, and complete the check for 2. Spend longer whenever you need.</p><p>Floating-point calculations can differ slightly after reordering. Compare an optimized implementation with a trusted reference using documented absolute and relative tolerances, and reject NaN when it is not expected. Never choose an enormous tolerance merely to make a test pass. Test zero, negative values, large magnitudes, and the application’s normal range.</p><p><strong>Practice and evidence:</strong> Predict the example or draw its state transitions. Change one assumption, test the boundary, and record the actual result. Explain what this small model does not establish about a production system.</p><pre><code class="language-python">import math
+actual=0.1+0.2
+print(math.isclose(actual,0.3,rel_tol=1e-9,abs_tol=1e-12))
+</code></pre><p><strong>Expected result:</strong> <code>True</code></p><h3>Practice and evidence</h3><p>Keep three notes: your prediction, the actual result, and why they match or differ. The completion check grades understanding; the practice task builds skill.</p><p><a href="https://git.agilaos.com/matt/open_agx">Language reference: git.agilaos.com</a></p><p><a href="https://onnx.ai/onnx/intro/">Language reference: onnx.ai</a></p></html>

en/expanded-25-1/lesson-3/main.py

@@ -0,0 +1,3 @@
+import math
+actual=0.1+0.2
+print(math.isclose(actual,0.3,rel_tol=1e-9,abs_tol=1e-12))