How DMI works

No self-reported progress. No participation trophies. Your real GitHub work is reviewed every day, graded automatically, and published. Here's the exact system.

One week in DMI

The same loop, every week, for 14 weeks

Saturday — Live session day

Pravin opens with the helicopter view of the week's topic and teaches the key concepts. Then 3–4 assignments are taught live with theory and hands-on demos — followed by peer reviews and Champion of the Week.

Sunday — Self-study day

Re-watch the session, take notes, revise the concepts, and start your assignment in your own GitHub fork of the DMI template repo.

Monday–Thursday — Execution days

Build the assignments in your fork — real files, real commits, your repo. Live assignment help from co-mentors in the evenings, and doubts cleared in the Discord community.

Friday — Deadline day

Push your completed work to GitHub by end of day. Your work lives in public — that's the point, it becomes evidence.

Every day — Automated review

The review agent checks every file of your work against the week's checklist and updates your public badge page and the leaderboard — daily, not just at week's end.

Every week — Share what you built

A weekly LinkedIn post and blog post about what you built earn points too — and build your public presence.

Self-Paced students follow the same loop at their own pace — same assignments, same daily grading, no live sessions.

How points are earned

Graded automatically, the same rules for everyone

WhatPoints
Each assignment file passing all checks20
Weekly LinkedIn post (URL in your progress table)10
Weekly blog post (must load + credit DMI)30
Live-session attendance (per slot: Morning · Mid · Evening)10 each

Live-session attendance applies to the live cohort and counts from Week 2 onward.

Real work only. Placeholder text scores 0. Copied or untouched files score 0. Answers under 50 real words score 0.
No pre-filling ahead. Grading is capped at the cohort's current week — you can't stack up points by pasting future weeks early.
Every run is kept. Score history is stored permanently, so any result can be looked up — not argued about.

The exact rules, in full

The same rules run against every student, every week — nothing here is decided case by case

One assignment file20 pts · pass or fail

A file fails automatically if any of these is true:

  • It isn't at the exact path the template expects — same week folder name, same assignment-*.md filename. There is no fallback matching, so keep your fork synced with the template.
  • It is byte-for-byte identical to a file in the template repo — an untouched stub, or a copied answer key.
  • It still contains template placeholder text: Add your answer here, Add your screenshot here., Add your URL here, or Add your post content here.
  • It has fewer than 50 words of your own writing, counted after stripping code blocks, HTML comments, headings and table formatting.

Otherwise it passes and is worth 20 points. There is no partial credit inside a file.

Weekly LinkedIn post10 pts

Read from the Weekly Progress table in your fork's root README.md — your own table, never inferred from anywhere else. The URL must begin with https://www.linkedin.com/posts/. A profile link or a shortened URL scores 0.

Weekly blog post30 pts

Same table. All three must hold:

  • The URL loads.
  • The page has at least 200 words.
  • The page contains a real link to your own badge page — dmi.pravinmishra.com/s/<your-username>.html.

A genuine, well-written post that links some other DMI page, or only mentions DMI in text, scores 0. The requirement is that every weekly post sends real traffic to your own graded work.

Live-session attendance10 pts per slot

Morning, Mid and Evening slots on Saturday, taken from the team's attendance record — not derived from your GitHub activity. Applies to the live cohort, and counts from Week 2 onward — weeks 00 and 01 carry no attendance component at all.

How a week adds up

Week total = assignment points + LinkedIn + blog + attendance. Each week's maximum depends on how many assignment files that week has, so totals aren't directly comparable week to week. Your grand total is the sum across every graded week.

Once earned, it stays earned

A week whose assignments have fully passed, and a blog post that has been verified, are locked — they are never re-checked. LinkedIn and attendance are recomputed on every run, so a corrected or backfilled attendance record always takes effect. If a rule changes later, credit you already earned under the older rule is not taken back.

What these checks cover

The agent enforces that the work is real, original and on schedule: present at the right path, written by you, not a placeholder, and substantial enough to be a genuine answer. It does not judge whether your answer is technically correct.

A public profile that proves it

Every student gets a public badge page: week-by-week scores, passing checks, LinkedIn and blog links — updated automatically after every review run. This page is public. Put it in your CV. Hiring managers can click every claim.

And a certificate, once you finish. Complete all 14 weeks — every week reading Complete on your badge page, 14 of 14 — and you receive a completion certificate co-issued by The CloudAdvisory and Pravin Mishra. That is the entire bar: the same daily automated grading everyone else is measured by, with no separate exam and no discretionary review. Self-Paced works the same way, across 14 sections instead of 14 weeks.

The certificate links back to your badge page — so it isn't a claim anyone has to take on trust. Whoever you show it to can open the graded work behind it.

See the live leaderboard → Click any student's name to open their real badge page.

Ready to build your proof?

Start the Self-Paced track today, or join the waitlist for the next live cohort.

Start today — Self-Paced Track → Join the Cohort 4 Waiting List