Zero to Hired: from a blank GitHub to an interview-ready portfolio in six months.

Sample page: the Pick-your-track guide mapping target roles to their 2-3 projects.

Sample page: the LinkedIn playbook that reads as engineering, not marketing.

Sample page: a RAG + eval-harness project with a free stack and a week-by-week plan.

You can code. Your GitHub doesn’t prove it, and you don’t know which projects matter for FDE, AI-ML, or data roles. Zero to Hired is the fix: 16 production-grade projects, a git + LinkedIn playbook, and the resume bullets each one produces.

Get the playbook · $50 $45

Instant access · 7-day refund

The problem.

The gap

Every company building with AI asks the same question: how did you know it worked? Here’s what they find when they open your GitHub.

career-project/

main.py
model.pkl
README.md
requirements.txt

$ git log --oneline
a3f9c21 final version

“I’ll add tests later”nothing proves it runs
“the README explains it”README says “TODO”
“five projects show range”zero depth, nothing to defend for ten minutes straight
“one clean commit is fine”no history, no proof you built it yourself

Zero to Hired replaces this: 2-3 deep projects, real git history, numbers you can defend.

What's inside.

Inside

52 pages, five parts. The literal list first, then what each one actually does.

Pick-your-track guide3 tracks, mapped to real roles
16 project specs30 of the 52 pages
Free-tier stack guidecommercial-grade tiers, and their limits
Three playbooksgit, DevOps, LinkedIn
Companion zipfile[Core]skeletons, tracker, bullet bank, Discord

The sixteen specs are your options, not your workload. Understand what the role you want actually screens for, then build the three projects that prove it. Depth is the whole point. Here is how that maps for three sample roles:

Forward-Deployed Engineer
- Approval-gated support agent
- Customer data integration service
AI-ML Engineer
- RAG system with an eval harness
- Drift-monitored retraining pipeline
Data Engineer
- Real-time clickstream pipeline
- Warehouse model + analytics API

Once you have picked those projects, the playbook hands you the same four-step build order for every one of them. Follow it end to end and the project finishes as something you can talk about, not just something that runs:

gitcommit-discipline ladder: your history becomes the evidence

What to build

The scope, and the hiring signal it sends.

The build plan

Week by week, to a working system.

The interview

The questions this project prepares you to answer.

The bullets

What goes on the resume when it ships.

DevOpsnetworking + deployment depth

LinkedInposts that read as engineering

The $0 stack

Groqinference
GeminiLLM + embeddings
ColabGPU
Oracle Cloudalways-free ARM
SupabasePostgres + auth

Run that four times and you have a portfolio. To make week one about building instead of scaffolding, the companion zip ships every file the specs assume you already have:

companion.zip

repo-skeleton/
README.md the one a recruiter reads first
design-doc.md decisions, tradeoffs, numbers
.github/workflows/ci.yml tests green on every push
CLAUDE.md agent context, preconfigured
progress-tracker.md 6 months, week by week
resume-bullet-bank.md phrasing per project
discord-invite.url feedback and accountability

GitHub
Claude
Discord

Why trust this.

Proof

You are standing in it. My portfolio is what this method produces, so every claim below is something you can go and check right now.

Seven case studiesRole-targeted projects, each written up so that even a private repo still reads as credible to someone who cannot see the code.
yelp-ml-platform (opens in new tab)Public repo. 7M reviews, ETL at ~462K rows/sec, 86.3% sentiment accuracy, p99 0.11 ms serving. Every headline number has a committed benchmark output behind it.
cohors (opens in new tab)Open source. Read the commit history instead of taking my word about commit discipline.

No hustle, no inflated numbers. The playbook tells you to publish only numbers you can defend for ten minutes, because one hard question in the comments undoes months of credibility. I hold myself to the same rule here.

Pricing.

Core

$50$45

One payment · lifetime updates · 10% launch discount, pre-applied

All 16 project specs, with week-by-week build plans
The pick-your-track guide
The free-tier stack guide
git, DevOps and LinkedIn playbooks
Companion zip: repo skeletons, tracker, Resume Bullet Bank
Buyers' Discord
Lifetime updates

Get the playbook · $45

Launch price $45 (10% off), then $50. A live cohort opens later; Core buyers get it at a discount, and everything in Core carries over.

Instant accessLifetime updates7-day money-back guarantee

FAQ.

Common questions

What do I get?

A 52-page PDF and the companion zip. Gumroad delivers both instantly at checkout.

Do I need a GPU or a budget?

No. Every project runs on free tiers.

Solo, or do I need a partner?

Solo. An appendix covers the two-person workflow if you find one.

Is this just a list of ideas?

No. Week-by-week plans, repo skeletons, a tracker, and the git and LinkedIn playbooks.

I'm not in the US.

Fine, none of it is US-specific. Price is USD; your card converts at checkout.

Is there a discount?

The $45 launch price is 10% off and already applied. After that it's $50.

Will it go stale?

Lifetime updates. The stack moves, so does the playbook.

Guarantee.

7 days

If the playbook doesn’t hand you a plan you are actually excited to start, you get your money back. No form, no justification, no “what didn’t you like”.

From your Gumroad receiptThe refund request is a button in the emailed receipt. You never have to talk to me.
Email mebhavsarrushir@gmail.com. One line is enough; I refund it and we are done.

Six months from now your GitHub either proves something, or it doesn’t.

It is the same six months either way. The only question is whether you spend them building the right three things.