AI Product · Creator Operations · Technical Content

FarhadMojirsheibani

I direct AI to build. I run what it ships.

Southern California, open to fully remote. I take AI ideas from a raw concept to a live, integrated system by directing AI-assisted workflows through UX, integrations, testing, and iteration. Two live systems, one channel, and a habit of finishing what I start.

Hire me → Download résumé ↓ See the work
GitHub YouTube Apple Music
Build case studies
5
One private AI tool, one public pre-launch platform, one personal offline-first PWA, one prototype, and one Python learning series.
Integrated services
~12
Claude, Cloudflare, Supabase, Voyage, Google OAuth, Gmail, YouTube, and more, wired end to end by me.
Videos published
840+
A tech channel I research, write, produce, and analyze on my own.
Location
SoCal
Based in Redlands, CA. Open to fully remote roles across the US.

Selected Work

Five systems I took from idea to shipped.

01   05

I own the parts that decide whether something is actually good: the idea and UX, the data and integration decisions, the testing, and the iteration. Where something is private or in progress, it says so.

01ProductLive · In use Private tool

Daryl, Private AI Creator Operations Assistant

I built Daryl to run the business side of my YouTube channel: live channel stats, sponsorship outreach drafts that always route through me before sending, persistent vector memory that learns me over time, and live research on the topics I make videos about.

PayoffHours of channel admin collapse into reviewed drafts, so my creative time goes back into making videos.
Role
Product direction, UX design, integrations, testing
Stack
Cloudflare Workers · Claude API · Supabase + pgvector · Voyage embeddings · Google OAuth · Gmail & YouTube APIs
Does
Live YouTube stats, scheduled outreach drafts held for approval, vector memory, live web research, voice input
Proof
Dated build record →

It prepares and drafts. I review and send. It never sends email on its own, and no credentials or private data are exposed here. Daryl itself stays private; the demo you see here is the public window into it.

Read the full case study →
► Watch full demo
02Consumer productDeployed pre-launch fuel-savings platform Repo coming soon

GasElite, Public Pre-Launch Waitlist Platform

A subscription consumer product on its own domain. I built the waitlist front-end, wired the database, set up automated confirmation email, verified the sending domain, moved DNS to Cloudflare, and hardened deliverability. The waitlist is live at gaselite.online. The full product is in stealth.

PayoffI built the waitlist front-end, wired the database, set up automated confirmation email, verified the sending domain, moved DNS to Cloudflare, and hardened deliverability by hand.
Role
Product direction, brand, front-end, DevOps, security, deliverability
Stack
React 19 · TanStack Start SSR · Cloudflare (DNS, Redirect Rules, Email Routing) · D1 database · Resend
Does
Live waitlist with US and international phone validation, automated branded confirmation email, contact form, privacy and terms pages, admin dashboard
Proof
Dated build record →

The waitlist is public. The full product is in stealth and I do not describe its features here on purpose.

Read the full case study →
► Watch full demo
03Personal toolShipped · PWA Repo coming soon

Hair Recovery Command Center, Offline-First PWA

A personal daily-protocol tracker I built for my own use, shipped as an offline-first Progressive Web App. Today view, 28-day calendar, weekly plan, four-week progress check-ins with photo notes, and a safety dashboard. Installs to a phone home screen. Runs without a network.

PayoffProof that I ship polished consumer-grade tools end to end. Real design, real engineering, real use, in a single self-contained file.
Role
Product, UX, front-end, PWA engineering
Stack
Vanilla HTML, CSS, JavaScript · Service Worker · Web App Manifest · localStorage · iOS and Android home-screen install
Does
Daily tasks, calendar, weekly protocol, progress check-ins, safety dashboard, program settings, local data with export

A personal tool. Not medical advice. Prescription labels and clinician instructions always override the planner.

Read the full case study →
► Watch full demo
04ToolPrototype Repo coming soon

RemoteFit, Remote Role Finder & Application Tracker

I defined the fit logic, categorized 100 remote roles, tagged résumé variants, and built application-status tracking with direct employer-search handoffs. A scattered job hunt becomes one ranked, trackable view.

PayoffLess time lost hopping between tabs. Better-matched applications, tracked in one place.
Role
Product direction, UX, data curation
Data
100 manually curated roles, direct employer and job-board links
Does
Fit-ranking, role categorization, résumé-variant tagging, application-status tracking

A manually curated list, not a live scraped feed. Individual job links can expire and should be verified before applying.

Read the full case study →
► Watch full demo
05LearningTwo models · Python View on GitHub

F800 Diagnostic Series, Python + AI Integration

I built command-line diagnostics for capability scoring and system-health calculations. Model 001 integrates Gemini for the analysis step. Model 002 documents the iterative Python learning behind it. Two videos side by side because the story is the progression.

PayoffProof that I close skill gaps by shipping. Built during my hands-on Python learning progression on an older Intel MacBook: I model the workflow, write the code, wire up the Gemini integration, and document each iteration honestly.
Role
Personally written Python, Gemini integrated for analysis
Stack
Python, Gemini API
Does
Input-driven capability scoring, calculations, and an AI-powered analysis and feedback step

A learning project, not an autonomous AI system and not an enterprise product. The value is the practice, the iteration, and the honest write-up.

Read the full case study →
Model 001 ► Full
Model 002 ► Full

Creative Direction

Prompt writing and art direction, on Midjourney.

The prompts, the composition calls, the color choices, and the iteration are mine. The images themselves are generated by Midjourney. I show this as direction studies, not fan art, and never with copyrighted characters.

I approach image generation the same way I approach code: a clear intent, a controlled prompt, a shot review, an iteration. Direction is the leverage. The tool is the multiplier.

This is the same instinct I bring to AI product work. The judgment about what to make and whether it landed is mine. The model is just how I make it fast.

► Watch full demo

Content & Channel

A tech channel I run end to end.

I research, script, produce, publish, and analyze consumer-technology content across long-form and short-form. I use audience signals to refine topics, clarity, and packaging. The channel is proof I can execute in public, self-direct, and communicate a hard idea to a wide audience.

4,000+
Subscribers
840+
Videos published
@SheibaniOfficial
YouTube channel

Figures captured August 2026. Not a live count. Visit the channel →

91.8%
Average view retention
Top 5%
Of all creators globally
The math behind the claim

91.8% average view retention is my own verified channel figure. The "top 5% of all creators globally" positioning is a benchmark analysis I wrote called The Retention Frontier, triangulating six published 2026 retention studies (Virvid, Lenostube, Retensis, vidIQ, Loopex Digital, Shortimize). The platform average is 73% and the average of the most-viewed shorts on the platform is roughly 76%; sitting at 90%+ places the channel in the tail.

Read the full analysis →
► Watch full walkthrough
Also, on the side

Independent Music · King Sheibani

I release music independently under the name King Sheibani. Distributed through DistroKid and available on Apple Music. It is a creative outlet, not a business line, and I do not claim performance or revenue numbers here.

Listen on Apple Music →

Skills & Stack

What I bring on day one.

Grouped by what I actually reach for. Honest depth marked in the notes.

Architecture before syntax

I read AI as leverage. Systems are how I think.

I understand problem architecture before I have the programming vocabulary to express it. That is the mismatch that makes me useful in AI product roles. I do the modeling, the judgment, and the reviews. The model translates my architecture into code, and I verify every piece before it ships. That is why I have shipped five real systems — one private production tool, one public pre-launch platform, one offline-first PWA, one working prototype, and one Python learning series — while I am still building depth in the syntax underneath them.

AI & APIs

Directing AI to build real things

Claude APIGemini APIVoyage embeddingsAgentic tool loopsPrompt designRAGVector memoryVision

Daily hands-on. Built and shipped agentic systems with tool use, memory, and multi-service integration.

Backend & Infra

Set up, wired, and debugged live

Cloudflare WorkersDNS & CloudflareSupabase + pgvectorGoogle OAuthGmail APIYouTube Data APIREST APIsResend

Set up every one of these myself, including the hard parts (OAuth refresh flow, vector memory, email deliverability).

Product & UX

The parts that decide whether it is any good

Product directionUX designWorkflow designResearch synthesisOutput evaluationFunctional testingIteration

Every project here is a decision I made and defended. Systems that a real person can pick up and use, not demos that only work once.

Languages & Frontend

Where I write and where I direct

PythonJavaScriptTypeScriptReactHTML & CSSGit & GitHub

Python I write and read directly (F800 is my own code). JS, TS, React and web work I ship with AI as the builder while I own the architecture and the reviews.

Content & Media

Communicating the work in public

YouTube AnalyticsSponsorship outreachFinal Cut ProCapCutCanvaPublic writing

840+ published videos, real audience feedback loop, and a habit of turning a rough draft into a shipped piece.

Human

How I show up

EnglishFarsiSelf-directedFast learnerWrites wellShips

Independent tutor since 2015. Comfortable adapting explanations to the person in front of me, and comfortable being wrong out loud so I can get better fast.

About

How I work.

Education & credentials

California State University, Northridge · B.A. Child & Adolescent Development, Early Childhood option, expected December 2026.

Master of Science in Computer Science pathway · CSUN department-guided plan, target December 2028.

SoloLearn Python Developer certificate, August 2026 · CPR / AED / First Aid, valid through April 2028 · California BSIS Guard Card.

  • Product direction
  • UX design
  • AI workflow design
  • Prompt design
  • Integrations
  • Testing & iteration

I focus on AI product, creator operations, and technical-content work, with direct hands-on experience directing AI-assisted workflows from a raw concept to a live, integrated system. That means UX, integrations, testing, and iteration. Not one of those in isolation. All of them, until something ships.

I build tools that turn everyday friction into money saved and time returned: a private creator-operations console that runs the business side of my channel, a public fuel-savings platform that pays for itself the first time a driver skips one bad pump, a personal offline-first PWA that keeps a daily recovery protocol on rails, and Python diagnostics that integrate AI, while I run a technology channel with 840+ videos and 4,000+ subscribers. The channel is the honest proof of self-direction: I research it, script it, produce it, publish it, and read the numbers back to make the next one better.

I am direct about how I build. Python I write and read myself. Web and backend code I ship with AI as the builder while I own the architecture, the integrations, and the reviews. The judgment about what to build, what to cut, and whether it is trustworthy is mine, and I stand behind it.

Hire me

Let's build the next AI system together.

Southern California, open to fully remote. I am looking for AI product, applied-AI, creator-operations, and technical-content roles where the work is real and the impact is measurable.

Email me → Download résumé ↓

Email farhad.mojirsheibani.83@my.csun.edu  ·  GitHub @sheibaniyt-lang  ·  Channel @SheibaniOfficial