Priya's Thursday, two ways
8 lessons · 25 prompts
A day in the life of an AI Engineer: what changes with these prompts.
Track progress as a memberPriya, an AI engineer at a small logistics company.
Priya starts Thursday with a messy shipment-delay dataset and a deadline for a first model. She opens ChatGPT and types: Write Python to load this CSV, fill missing port codes, and split train and test sets. She adds her column names and asks for a few tests. The draft is not perfect, but it gives her a clean starting point.
After lunch, a training run stops with a CUDA out of memory error. She opens Claude and types: Here is the traceback and my batch size. What should I change first? Claude suggests gradient accumulation and a smaller batch. Priya tries both, and the run moves forward.
Later, she uses Gemini for two more lessons. For model training and tuning, she types: Help me plan a small hyperparameter search for learning rate and dropout, and explain how to read the learning curves. For monitoring and optimization, she types: Draft a checklist for drift alerts on arrival-time features and a query to compare last week to this week. She edits both to fit her team's dashboard.
By the end of the day, Priya has a working data script, a training run that no longer crashes, and a monitoring checklist. She uses the time won back to leave on time, walk her dog through the park, and read a few pages of a novel before bed.
Before
- Too many tabs, too many errors.
- Code copied, then fixed by hand.
- Training runs fail late in the day.
- Notes scattered across chats and docs.
After this course
- One prompt starts the data script.
- Errors explained before coffee goes cold.
- Training plans checked in a short chat.
- Monitoring checklist ready before stand-up.
What you'll learn
- Code drafting: Generate and refine Python for data handling, training, and tests.
- Debugging help: Read errors and fix Python, TensorFlow, and PyTorch issues.
- Explaining papers: Understand algorithms and math behind your models.
- Data preparation: Plan cleaning, exploration, and feature engineering code.
- Training and tuning: Design loops, plan searches, and read learning curves.
- Deployment support: Draft serving code, container configs, and diagnose failures.
- Monitoring models: Set up alerts, spot drift, and improve speed and resource use.
- Docs and sharing: Write model cards, explain experiments, and update teammates.
How this course works
- 8 lessonsOne task of your job each, from code drafting basics to documentation and collaboration.
- Ready-to-paste promptsCopy, fill in the parts in {{brackets}}, paste into ChatGPT, Claude or Gemini.
- Tick and completeTick the prompts you tried and mark each lesson complete.
- Get certifiedFinish and keep the prompts as your own library.