Choose your starting point · 01 / 15
Choose the job you are preparing for
My recommendation: target an applied AI engineer or software engineer building AI features. Learn enough software engineering to own an application, then show that you can evaluate, deploy and improve a useful model-powered feature.
This is my practical route through the field, built for this library. It complements the three other classes. It prepares you for application work; research scientist and model-training roles require a different, deeper path.
Know which job title you mean
| Role family | Typical work | Where this path fits |
|---|---|---|
| Applied / product AI engineer | Model APIs, retrieval, tools, evaluation, product features | Primary target |
| Backend / full-stack engineer, AI team | APIs, data, auth, deployment, AI integration | Strong adjacent target |
| Solutions / forward deployed engineer | Customer discovery, integrations, demonstrations, delivery | Good if you enjoy customers and can code |
| ML engineer / research engineer | Training, data pipelines, experiments, inference infrastructure | Add substantial ML and systems study |
What the research actually supports
I reviewed three employer postings as examples, not a representative labor-market survey. DeepIntent combines Python/API work with business workflows and asks for two or more years of development experience, including certain project experience. Schonfeld advertises multiple engineering levels and emphasizes platform quality, evaluation and observability. Anthropic’s customer-facing role asks for four or more years in technical roles and production LLM experience. DeepIntent ↗ · Schonfeld ↗ · Anthropic ↗.
My inference: a small, well-explained production project is a better organizing goal for this path than a long certificate list. A portfolio does not erase experience, education, location or work-authorization requirements. For a first engineering job, also look at junior backend roles, internships and supervised internal automation projects.
Your first 60 minutes
- Choose one role family and an industry you understand or can access.
- Collect 10 current postings where you could realistically work. Separate required skills from preferences; record experience and location constraints.
- Circle the five skills repeated most often. Pick a project that demonstrates three of them.
- Choose the beginner, engineer or vibe-coder route by the practical diagnostic, not by confidence alone.
You can name a target role, explain what it builds, and point to 10 relevant job descriptions. Recheck this sample monthly while studying.
Your lesson resources
Download these files to follow along and put the lesson into practice.