AI Engineering Core Skills — Model Selection and Code Generation

Led by Demis Hassabis Simulacrum

Selecting the right model for the task — benchmarks, leaderboards, evaluation methodology, and a practical case study in AI-powered code translation.

1. Module 1

   Benchmarks, Leaderboards and Selection Strategy

   Led by Demis Hassabis Simulacrum

   The question

   Model selection strategy and fundamentals · the Chinchilla scaling law (parameters vs training data) · understanding benchmarks (GPQA, MMLU-Pro, HLE) · limitations of benchmarks (data contamination, overfitting) · navigating leaderboards (Artificial ...

   Outcome

   Demonstrates engineering competence in benchmarks, leaderboards and selection strategy.

   Sub-units

   1. ○ 1.1 Benchmarks, Leaderboards and Selection Strategy Open this unit → ✓ ↻
   Begin the course → Print certificate of completion
2. Module 2

   Code Generation — A Model Selection Case Study

   Led by Demis Hassabis Simulacrum

   The question

   Selecting models for code generation tasks · Python to C++ translation with frontier models (GPT-5, Claude, Gemini, Grok) · measuring performance speedups · open-source models for code generation (Qwen, DeepSeek, Ollama) · building a Gradio UI for co...

   Outcome

   Demonstrates engineering competence in code generation — a model selection case study.

   Sub-units

   1. ○ 2.2 Code Generation — A Model Selection Case Study Open this unit → ✓ ↻
   Open this module → Print certificate of completion