Frontmatter Daniel Godoy, Author of Deep Learning with PyTorch Step-by-Step: A Beginner's Guide
Mar 31, 2022
Daniel Godoy, a Berlin-based data scientist and author, shares his journey from Brazil’s early programming days to a successful career in data science. He recounts how a stochastic simulation paper changed his career trajectory and led him to teach in fintech. The COVID lockdown inspired him to write his book on PyTorch, where he emphasizes the importance of formal education in programming. The discussion also covers deep learning fundamentals, the risks of biased datasets, and the complexities of self-driving cars, showcasing his passion for teaching and knowledge sharing.
AI Snips
Chapters
Books
Transcript
Struggle First, Master Later
- Early MATLAB work on vectorized stochastic simulations provided foundations that translated directly to modern NumPy and Python workflows.
- Struggling with fundamentals first can make later tools much easier to master.
Use Monte Carlo To Capture Uncertainty
- Use Monte Carlo (stochastic) simulations to model uncertainty by sampling many random scenarios and averaging results.
- Analyze distributions (e.g., percentiles) not just averages to understand risk and variability.
Limits Of Economic Assumptions
- Economics behaves like a science but relies on unrealistic assumptions (e.g., ergodicity) that can mislead practical decisions.
- Real-world single-agent constraints mean expected-value logic can be dangerously inapplicable.
