FIELDGUIDE
THE PRACTICAL AI/ML PATH

Less planning.
More building.

A map from “I can code a little” to “I can build, explain, evaluate, and ship an AI project.” One phase at a time. No prerequisite-marathon required.

Explore the phases
Curated around free, reputable resources. You do not need to finish every link.
◎BUILD
LOOP
▦DATAfind patterns
⌘MODELtest a hypothesis
↗SHIPmake it useful
LEARN → TRY → EXPLAIN → REPEAT
YOUR JOURNEY0% complete
0 of 0 milestones0 / 9 phases
YOUR NEXT SMALL WIN

Set up your learning workspace

Do one small task today. Progress comes from finished attempts, not perfect plans.

THE 25-MINUTE RULE

Make starting ridiculously easy.

One tab. One task. Stop when the timer ends—or continue if you're in flow.

25:00
THE ROUTE

Your path, in phases.

Follow the order, but don't wait for perfect understanding. Each phase ends with something you can demonstrate.

Core path Optional later
ANTI-OVERWHELM SYSTEM

Rules that keep you moving.

Use these when your attention drops, a topic feels hard, or you want to start another course.

01

One main resource

Pick the core resource listed in your current phase. Use extras only when something is unclear. No collecting courses.

02

Copy, then change

Following an example is okay. Next, change an input or method, predict the result, run it, and explain what happened.

03

Build before you feel ready

Start the tiny project early. Google errors, read docs, and ask for help—but write down what you learned to fix.

04

No streak guilt

Missing days is normal. There is no streak to protect. Return by doing the next 10-minute action, not by restarting.

Placement season is still active.

If you're applying to general software roles, keep DSA, CS fundamentals, and interview practice as the main track. Use AI/ML as a focused second track. This roadmap is designed to progress in short sessions.

KEEP IT SIMPLE

Your toolkit.

Don't install everything today. Use tools when a phase needs them.

Python + notebooks

Experiment in small cells; keep work reproducible.

Jupyter ↗

Data work

Tables, array operations, and quick visual checks.

pandas tutorials ↗

Classical ML

Start with reliable, well-documented algorithms.

scikit-learn ↗

Version control

Save checkpoints and show what you actually built.

GitHub Skills ↗

A map, not a promise.

Finishing checkboxes is not the same as mastering a topic. Use the exit checks to test yourself honestly. Resource links can change; if one breaks, use the official documentation linked nearby. Most listed core resources are free, but external platforms may change access or terms.