Machine Learning with NO Tech Background?
Marina Wyss’s Guide to Breaking In
Ever feel like you missed your shot at tech because you didn’t come from a traditional computer science background? This week’s guest will make you rethink that entirely.
We sat down with Marina Wyss, Applied Scientist at Twitch, creator of the "Gratitude Driven" YouTube channel, and self-taught machine learning applied scientist, to unpack her inspiring and highly tactical journey into the world of ML. Spoiler: she didn’t start with Python. She started with political science and managing a jewelry business.
But through deliberate self-study, cold emailing, strategic upskilling, and a deeply human approach to productivity (think gratitude > grind), Marina broke into big tech - without a CS degree or a bootcamp.
In this episode, you’ll learn:
How to reverse-engineer your way into ML jobs by building “luck surface area”
The core skills Marina recommends for beginners (hint: SQL is underrated)
Why doing Kaggle projects isn’t enough - and what to build instead
How AI tools like ChatGPT can accelerate your learning if used well
Her take on career growth, promotions, and why communication trumps code
If you're feeling stuck, overwhelmed by tech content overload, or unsure of your first (or next) step, this episode delivers a clear, motivating path forward.
Key Takeaways:
You don’t need a tech background to work in ML.
Marina’s journey from public policy to applied ML proves that curiosity, persistence, and self-study can bridge the gap.Learning should be problem-driven.
Instead of mastering every tool, Marina recommends learning just enough to start building, then tackling problems as they arise. This reduces overwhelm and accelerates depth.AI is a learning assistant, not a shortcut.
Marina shares how she uses ChatGPT to quiz herself, debug errors, and even design learning roadmaps - but warns it can still hallucinate confidently.Real-world projects beat polished tutorials.
Building something you care about (fitness app, personal dashboard, etc.) keeps you engaged and demonstrates initiative to recruiters.Communication is the most underrated superpower.
From interviews to networking to collaboration, your ability to clearly convey ideas often matters more than your ability to write perfect code.
Resources Mentioned:
Designing Machine Learning Systems - Chip Huyen
AI Engineering - Chip Huyen
Software Engineering for Data Scientists - Catherine Nelson
Machine Learning Specialization (Andrew Ng, Coursera)
Deep Learning Specialization (deeplearning.ai)
3Blue1Brown - Essence of Linear Algebra & Calculus
Follow us on LinkedIn (tell us Substack sent you!):
#machinelearning #datascience #careertransition #selftaught #chatgpt #gratitudedriven #nontech #mlprojects #deeplearning #python #DataNeighborPodcast