Exact Science
Get started
ProgrammingWeekly course

Practical AI

Use AI tools in real workflows with a clear mental model of what the systems are doing, so you can ship work without guessing.

About this course

A project-based course for early-career professionals who want AI in their workflow as a reliable system: tools and implementation first, with enough of how it works that you never feel lost.

Over eight weeks you will ship three working systems: a Communication OS, a Research & Decision Pack, and one Live Workflow used on real work with a measured gain. You will build a clear mental model of capabilities, context, and failure modes, then put it to work.

Format

8 weekly sessions × 90 minutes, small group (up to 6). Homework applies that week's pattern to your actual job tasks.

What you leave with

three shipped systems, a personal operating rhythm, and a 30-day maintenance plan for the workflow that matters most.

This course is not

becoming an ML engineer, chasing every model release, or “AI strategy” theatre without shipped workflows.

Weekly plan

Project-based

Project map

  1. 1

    Communication OS

    (weeks 1–3): email, docs, and meeting templates that preserve your voice and include human checkpoints.
  2. 2

    Research & Decision Pack

    (weeks 4–5): a repeatable briefing → memo pipeline with sources and verification steps.
  3. 3

    Live Workflow

    (weeks 6–8): one automated or agent-assisted job workflow used on real work, with an ROI note and 30-day maintenance plan.
1

Week 1

How it works enough (then audit your week)

Tools
one frontier chat model.
In class
capabilities vs limits; context and hallucination in workplace terms; map tasks into automate / accelerate / advise / avoid.
You ship
audit of your real week + project shortlist.
Project checkpoint
P1 start.

Teacher notes

Session goal: mental model before tooling. Demo failure: automating a low-value task. Exit: audited week and three candidate workflows.
2

Week 2

Briefing, verification, and unsafe domains

Tools
chat + checklist templates.
In class
what “context” means in practice; spec writing; second-opinion patterns; legal, HR, and client-fact risk.
You ship
verification checklist for your role.
Project checkpoint
P1 build.

Teacher notes

Session goal: briefs that survive bad inputs. Demo failure: fluent memo with a wrong client fact. Exit: checklist ready for live threads.
3

Week 3

Writing and communication OS

Tools
chat + email/docs.
In class
voice preservation; meeting notes → action items; templates with human checkpoints.
You ship
Communication OS v1 used on live threads.
Project checkpoint
P1 ship.

Teacher notes

Session goal: usable templates, not a style guide nobody opens. Demo failure: generic corporate voice. Exit: OS used on at least two real threads.
4

Week 4

Research and analysis

Tools
chat + sheets/docs.
In class
briefings, scans, structured outputs, decision memos; check one number yourself every time.
You ship
first decision memo from real inputs.
Project checkpoint
P2 start.

Teacher notes

Session goal: structure before prose. Demo failure: long summary with no decision. Exit: one memo a colleague could act on.
5

Week 5

Knowledge systems

Tools
Projects / files-as-context (Notion, Drive, or similar).
In class
personal wiki as RAG-lite; memory hygiene; keeping context clean.
You ship
Research & Decision Pack v1.
Project checkpoint
P2 ship.

Teacher notes

Session goal: reusable context, not chat archaeology. Demo failure: stale file that steers the model wrong. Exit: pack with source folder, memo template, and verification steps.
6

Week 6

Automation and light agents

Tools
Zapier / Make / n8n + LLM steps.
In class
scheduled digests, intake forms; when to stop and hire a developer.
You ship
workflow sketch with human checkpoints.
Project checkpoint
P3 start.

Teacher notes

Session goal: thin automation with brakes. Demo failure: fully autonomous loop on sensitive data. Exit: sketched Live Workflow with named checkpoints.
7

Week 7

Build for your role

Tools
role-track tools (ops, marketing, PM, or technical-adjacent / Cursor).
In class
deep work on one live job task end-to-end.
You ship
Live Workflow mid-build on a real job task.
Project checkpoint
P3 build.

Teacher notes

Session goal: real work only. Demo failure: demo data that hides edge cases. Exit: mid-build running on a real input from that week.
8

Week 8

Demo day

Tools
full stack of your three projects.
In class
live demo; ROI note (time saved / quality improved / risk reduced); maintenance checklist.
You ship
Live Workflow shipped + 30-day plan.
Project checkpoint
P3 ship.

Teacher notes

Session goal: prove the system on live work. Demo failure: cannot state the human checkpoint. Exit: three systems shipped; ROI note and 30-day plan written.

We'll call or email you back

Mon–Sun, 8am–9pm

Questions?

Get in touch

Ask about this course or anything else. We will help you choose what works best for you.