P01 / ECONOMICS
NYU · Year 3
Economics major, CS + Math minors. [Farida: coursework or GPA highlight]
SPEAKS CAC, LTV, GPU-HOUR MARGIN.
00 / 05 — OPERATOR
FARIDA
SHADY
CANDIDATE — AI PARTNERSHIPS & GROWTH INTERN
INFERENCE.AI · REDWOOD CITY, CA
Economics × CS × PM. I turn builder communities into compute customers.
NYU · NEW YORK · [Farida: email] · [Farida: LinkedIn URL]
EDIT Pink dashed fields are editable — click, type, it saves in this browser.
— PROOF OF WORK
P02 / COMMUNITY
[Farida: club name]
[Farida: role, e.g. President] — [Farida: members grown X→Y]
RECRUITED. RETAINED. SHIPPED.
P03 / EVENTS
[Farida: events run + total attendance]
Flagship: [Farida: flagship event example]
TAB 03'S BUILD-NIGHT MOTION, POINTED AT GPUS.
P04 / PRODUCT
[Farida: product or feature shipped]
Outcome: [Farida: PM outcome metric]
SPEC → SHIP → MEASURE. SAME LOOP AS GROWTH.
— WHY INFERENCE.AI, WHY ME
- The role is one funnel. Academy credits → Ghost VM → Maestro usage → paid Engine customer.
- Running it takes all three. Economics to price CAC against GPU-hour margin. CS to demo "first token in 15 minutes" without an SE. PM to run the loop weekly.
- This dashboard is the application. Tab 02: my campus, mapped. Tab 01: the sales asset. Tabs 03–05: how it scales past me.
— EXECUTION MEMO
MEMO BGOS-001 · STATUS: COMMITTED
First 30 days — NYU beachhead + one community channel
T0 = [Farida: start date]
TARGETS, NOT HISTORY — EVERY NUMBER BELOW IS A COMMITMENT.
W1
Map & instrument
Validate the Tab 02 list on the ground. Stand up pipeline tracking. Credit offer approved internally. 40 emails out by Friday.
KPI → 25 TARGETS VALIDATED · 40 OUTREACH SENT · OFFER APPROVED
W2
First contact
Wave two + follow-ups. Officer meetings booked. Two club co-hosts and a room locked. One builder Discord opened (Tab 05).
KPI → 100 CUMULATIVE OUTREACH · 8 MEETINGS HELD · 2 EVENTS BOOKED
W3
First build night
"Ship an agent before midnight." Ghost VM warm in ~60s, first token in 15 minutes. Activation happens in the room.
KPI → 40 ATTENDEES · 30 GHOST VMS ACTIVATED · 200 ENGINE GPU-HOURS
W4
Convert & systematize
48-hour usage follow-ups. Weekly office hours. ROI calc (Tab 01) preloaded for top builders. Playbook v1.1 ships with real NYU data.
KPI → 60 CUMULATIVE VMS · 500 GPU-HOURS · 5 PAID ENGINE CONVERSIONS
DAY-30 SCOREBOARD — COMMITTED TARGETS
150OUTREACH SENT
4EVENTS BOOKED
60GHOST VMS LIVE
500ENGINE GPU-HRS
5PAID CONVERSIONS
01 / 05 — BUILDER ROI CALCULATOR
Your workload. Both bills.
Engine bare metal at wholesale vs. a hyperscaler on-demand baseline. Pick a preset or dial in your own.
⚠ ILLUSTRATIVE RATES — editable constants in
app.js, not official pricing. Swap in live Engine list prices + a current hyperscaler quote before customer use.
$0
SAVED ON ENGINE
0%
CHEAPER
0
EXTRA GPU-HRS THE SAVINGS BUY
WHY THIS TAB EXISTS — THE ARTIFACT YOU HAND A CLUB OFFICER OR FOUNDER: THEIR EXACT WORKLOAD, BOTH BILLS. MAESTRO CAPS THE BILL IN PROD; ENGINE INFINIBAND SCALES THE MATH TO 4,096 CARDS.
02 / 05 — NYU BUILDER PIPELINE
The beachhead, mapped.
Every NYU org that burns — or should burn — GPU, with the wedge to reach it. Stage pills are editable and persist locally.
⚠ TYPED FROM NYU'S PUBLIC SURFACE — validated on the ground in Week 1. Contact paths are role-based; no invented emails.
| Org / Target | Type | Why they need compute | Wedge | Contact path | Stage |
|---|
— READY-TO-SEND OUTREACH
Named build. Concrete offer. One ask. Copy, fill the brackets, send.
03 / 05 — CAMPUS GTM PLAYBOOK
One system. Any campus.
NYU is the prototype. This is the manual a rep at any school runs: who to target, what to say, how to run the night, what to measure.
ICP — who actually burns GPU
TIER A — PURSUE FIRST
- AI/ML clubs with project teams — hit Colab limits every semester.
- Hackathon organizers — compute is a sponsorship line item they already shop.
- Course staff, GPU-heavy classes — one pilot = a roster of activations.
- Labs with cluster-queue pain — burst to Engine at deadlines.
- Student startups — seed money leaking into on-demand cloud.
TIER B — OPPORTUNISTIC
- Data/analytics clubs (Academy fit) · robotics (sim) · game dev (render) · quant (backtests).
DISQUALIFY
- No shipped output.
- Cash-only asks — the offer is credits + an event, never money.
SIGNALS → SHIPS PROJECTS? · GITHUB FOOTPRINT? · RECURRING EVENTS? · PUBLIC COLAB/QUEUE COMPLAINTS?
Outreach — cadence & rules
| Day | Touch | Content |
|---|---|---|
| D0 | Cold email + same-day DM | Template A. Names their last project. Ask: 20 minutes. |
| D3 | Follow-up | Template B. One line, restates the offer, semester deadline. |
| D7 | Value bump | ROI calc (Tab 01) preloaded with their workload. |
| D14 | Last call → recycle | Park for next semester. Officers turn over; the list compounds. |
- Name a concrete build — never "explore a partnership."
- Offer = credits + an event. Activation happens at events, not in inboxes.
- One owner per target. Stage updated same day.
Event motion — credits → build night → Engine
T-MINUS 2 WEEKS
- Club co-host locked: their room, our compute.
- Registration live; Academy codes per registrant.
- Theme: "Ship an agent before midnight."
NIGHT OF
- 0:00 — 5-min intro. Agents. GPUs. Inference.
- 0:05 — everyone claims a Ghost VM (~60s warm) + credits.
- 0:15 — first token, everyone.
- 0:20 — build sprint, floaters helping.
- END — 60-second demos; winner gets an Engine credit pack.
T-PLUS 48 HOURS
- Per-builder usage stats: "you burned 11 GPU-hours — here's both bills."
- Top 10% get a 1:1 + Engine offer.
- Everyone else: weekly office hours. The retention loop.
Funnel + KPIs
IDENTIFIED → CONTACTED → MEETING → ACTIVATED → HABIT (WEEK-2 RETURN) → PAID (ENGINE $)
| Step | Benchmark* | Lever if below |
|---|---|---|
| Contacted → Meeting | 15–25% | Sharper personalization; warm member intro. |
| Signup → Show | ~60% | Day-of reminders; co-host pushes their channels. |
| Attendee → Activated | 75%+ | Cut setup friction; more floaters. |
| Activated → Week-2 return | ~30% | Office hours, challenges, leaderboard seasons. |
| Returning → Paid | 15–25% | ROI calc with real usage; Maestro caps remove fear. |
*DIRECTIONAL — REPLACE WITH COHORT DATA AFTER EVENT #1. NORTH STAR: WEEKLY ACTIVE GHOST VMS PER CAMPUS. REVENUE: ENGINE GPU-HOURS.
Rep scorecard — owed weekly
| Metric | Weekly target* | Notes |
|---|---|---|
| New outreach | 40 | Personalized; Phase 01 cadence. |
| Meetings held | 5 | Officers, organizers, course staff, founders. |
| Events run | 0.5 | One build night per two weeks, minimum. |
| Ghost VM activations | 25 | Counted at the event, verified in the dashboard. |
| Engine GPU-hours | 250 | The number the business feels. |
| Paid conversions | 2 | Lagging — don't force before Habit. |
| Pipeline hygiene | 100% | Stage updated same day, no exceptions. |
*DIRECTIONAL — CALIBRATE AFTER THE FIRST FULL NYU CYCLE.
04 / 05 — TOP 25 BUILDER CAMPUSES
Where the playbook goes next.
Ranked by builder density — click any header to re-sort. NYU is the beachhead regardless of rank: operator on the ground.
⚠ CANDIDATE'S DIRECTIONAL MODEL, NOT OFFICIAL RESEARCH. Blend: ML research 30 · hackathons 25 · AI clubs 20 · startups 15 · GPU-access gap 10.
05 / 05 — DEVELOPER COMMUNITY GRID
Where builders already cluster.
Campus is the wedge, not the ceiling. Each card: a watering hole, the product that belongs there, the play to run.