Chai Visual

AI Roadmap

Tell it what you want to learn.
Get the whole syllabus.

Not a template. Every roadmap is drafted for your subject, your starting point and the minutes you actually have, ordered so nothing depends on something you haven’t learnt yet, then served one honest list each morning.

Create an accountLog in

Four questions · about half a minute · included in the paid plan

Today · Monday 3 Aug

4 to go

1 of 5
DSA for interviewsTwo pointers · day 15
SolveContainer With Most Water
Solve3Sum
JavaScript for frontend workAsync and the event loop · day 6
ReadMicrotasks vs macrotasks, and why a promise wins
PractiseRewrite a callback chain with async/await
PractiseCancel an in-flight fetch with AbortController
DSA for interviews
M
T
W
T
F
S
S
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
done1 missed
62% done
34 of 55 tasks
11
days left
3
in backlog

Any subject

Ask for anything. It sizes itself.

A language, a tool, a single data structure, an entire role. Type it in your own words. The time you give it is a ceiling, not a quota: it takes what the subject needs and then stops.

DSA for interviewsJavaScriptSystem designDockerReactLinked listsSpring BootSQLKubernetesRustGoNetworking…or yours
“Linked lists”
7 days
4 chapters · 20 tasks

One data structure. Covered, then finished.

“JavaScript”
30 days
5 chapters · 89 tasks

A language, from values and scope to the DOM.

“DevOps”
42 days
6 chapters · 126 tasks

A whole role. Linux first, Kubernetes last.

All three were asked for with six weeks available.

Under the hood

What the AI is actually doing

Four jobs, in order. Get the first one wrong and the rest do not matter.

  1. 01

    Drafting a contents page, not a task list

    It settles the chapters first and orders them by dependency, the shell before containers, containers before orchestration. A plan in the wrong order is useless however good the individual days are.

  2. 02

    Sizing the subject honestly

    Ask for one data structure and you get a tight week; ask for a whole role and it uses every day you offered.

  3. 03

    Reaching into our own problem library

    Anything algorithmic is built from problems that already live here, so a task opens on the page where you watch the algorithm move, not a link to somewhere else.

  4. 04

    Writing tasks you can start

    “Learn Docker” is a heading. “Write a Dockerfile for a Node app and cut its image size with a multi-stage build” is something you can begin now.

Life happens

Miss a day and nothing breaks.

Unfinished work becomes backlog and waits, grouped by the day it was meant to happen. Nothing is rescheduled, nothing expires, and the plan keeps the shape you agreed to.

Your midnight, not a server’s

Days roll over at midnight India time, so the plan says the same thing whether you open it in Delhi or on a laptop three time zones away.

You’re not doing this alone

A live count of what everyone else has been doing sits beside your plan. Counts only: no names, no leaderboard.

In the last hour

  • problems solved31
  • answers checked84
  • tasks ticked off57
  • learners joined12

across everyone learning here

Built in about half a minute.

Four questions, and you read the whole plan before it starts. Delete what you already know, add what you want.

Create an accountLog in