I turn raw ideas into AI-powered products, visual systems and working experiences.

JNASH graffiti logo

Creative Technologist, AI Product Designer, Brand Developer and Founder. I shape products, identities and cinematic stories, direct AI-assisted execution, and refine every detail until the experience behaves.

01AI PRODUCT DIRECTION
IdeaWorkflowPromptPrototypeTestShip
STATUSIndependent product creatorEvidence-led portfolio. Sensitive data removed.

Selected Work

Four case studies chosen from the local project audit.

The selection favors original product thinking, AI workflows, visual systems, automation and real problem-solving. Trading work is included, but it does not define the whole portfolio.

01IF

In development

Ink Flow

Studio operations usually scatter across messages, calendars, payment links and notebooks. Ink Flow frames that mess as a calm command center.

Role
Conceived and directed the product, defined the SaaS workflow, shaped the studio data model, directed AI-assisted development and verified the build path.
Tools
Next.js / TypeScript / Supabase / RLS / Tailwind / OpenAI-ready architecture
Open case study
02JV

Functional prototype

Jarvis Command Center

Generic chat hides context. Jarvis exposes mode, priority, cost, speech, weather shortcuts and tool-readiness as a visible machine.

Role
Directed the interaction model, interface language, runtime cards and AI-assisted build of the command center prototype.
Tools
React / TypeScript / Vite / Framer Motion / Three.js / OpenAI/Hermes-compatible API routes
Open case study
03SH

Static mobile prototype with backend-ready paths

Shadow

Most coaching apps collect habits without confronting the identity pattern underneath. Shadow turns behavior into logs, protocols and decisions.

Role
Conceived the product metaphor, defined the onboarding, directed the terminal-style mobile UI and shaped AI-assisted diagnostic flows.
Tools
React/Vite prototype / Supabase-ready memory model / OpenAI Edge Function path / Local fallback logic
Open case study
04PR

In development

Prop Account Rotator

Prop account management mixes emotional decisions, rule limits and platform friction. The rotator structures account eligibility and risk state.

Role
Defined product behavior, safety boundaries and workflow architecture while directing AI-assisted C# implementation and review.
Tools
C# / NinjaTrader / WPF dashboard / Risk engines / Simulation importers / License service structure
Open case study

Audit

Public-safe source scan.

FeatureInk Flow

Strongest SaaS product thinking and verified build evidence.

FeatureJarvis Command Center

Best AI interface and interaction-system proof.

FeatureShadow

Original coaching concept with clear AI-native workflow.

FeatureProp Account Rotator

Technically deep; included without letting trading dominate.

SupportHermes / Personal Agent

Useful agent and messaging-integration evidence.

SupportiTrade App

Trading coach surface; kept secondary.

SupportDuality Wellness

Strong visual/product-storytelling asset.

SupportPakito Nitro

Business tool and admin-workflow prototype.

SupportCuadrante Facil

Operational scheduling system in Spanish.

LabNasty Nash Arcade

Creative brand/audio/game experimentation.

Case Studies

Thinking, constraints, architecture and next moves.

01

In development / Documents/INK FLOW

Ink Flow

Context
A multi-tenant SaaS foundation for tattoo studios: onboarding, booking, CRM, dashboard operations, audit logs and future assistant workflows.
Problem
Studio operations usually scatter across messages, calendars, payment links and notebooks. Ink Flow frames that mess as a calm command center.
Constraints
Use only verified local evidence, keep secrets out, avoid invented metrics and preserve the project direction already present in the repository.
My role
Conceived and directed the product, defined the SaaS workflow, shaped the studio data model, directed AI-assisted development and verified the build path.
Product decisions
Built for Tattoo studios and artists who need one organized operating layer for leads, appointments, clients, deposits and follow-up. Core features: Tenant-safe studio model, Auth and onboarding, Dashboard shell, Requests, clients, calls and payments areas, Backup and recovery planning.
AI workflow
AI is intentionally staged: placeholders define assistant, voice and messaging pathways while keeping keys and customer data out of the client.
Visual process
Dark premium SaaS interface, compact navigation, restrained tattoo-studio tone and custom display typography decisions documented with font licensing checks.
Technical architecture
Static and server-rendered Next.js shell with Supabase helpers, RLS migrations, storage policies, audit tables and future integration slots.
Iterations
The inspected projects show an iterative pattern: mock first, shape the workflow, add integration placeholders, then verify build behavior before real credentials or customer data enter the system.
Failures or challenges
The safe portfolio version excludes unverified claims, private keys, customer/account information and any feature that exists only as a future idea.
Final solution
A multi-tenant SaaS foundation for tattoo studios: onboarding, booking, CRM, dashboard operations, audit logs and future assistant workflows.
Verified outcome
Verified from project docs: install, lint, typecheck and production build passed during the Phase 1 foundation.
What I would improve next
Connect real pilot studio data, confirm production language, add live AI assistant through server-side functions and finish payments/calendar integrations.
Sanitized / excluded
Excluded Supabase keys, service role details, storage credentials, payment setup and any pilot/customer data.
02

Functional prototype / Desktop/SHADOW/JARVIS

Jarvis Command Center

Context
A voice-forward AI command interface with lanes, runtime state, visual cards, model controls and a live conversation shell.
Problem
Generic chat hides context. Jarvis exposes mode, priority, cost, speech, weather shortcuts and tool-readiness as a visible machine.
Constraints
Use only verified local evidence, keep secrets out, avoid invented metrics and preserve the project direction already present in the repository.
My role
Directed the interaction model, interface language, runtime cards and AI-assisted build of the command center prototype.
Product decisions
Built for A founder/operator who wants an AI interface that can become a practical command layer across personal tools and business systems. Core features: Voice shortcut, Command palette, Model picker, Runtime cost panel, Weather shortcut, Streaming Hermes path, Visual settings.
AI workflow
Routes messages through OpenAI-compatible or Hermes endpoints, with lane instructions, streaming support and local weather handling.
Visual process
Orb-driven tactical HUD, terminal activity feed, configurable colors and a command-center layout built around visible system state.
Technical architecture
Vite client with API routes for chat, speech, weather, models and runtime status; local storage keeps visual and runtime preferences.
Iterations
The inspected projects show an iterative pattern: mock first, shape the workflow, add integration placeholders, then verify build behavior before real credentials or customer data enter the system.
Failures or challenges
The safe portfolio version excludes unverified claims, private keys, customer/account information and any feature that exists only as a future idea.
Final solution
A voice-forward AI command interface with lanes, runtime state, visual cards, model controls and a live conversation shell.
Verified outcome
Verified from source: the prototype contains live API routing, browser speech hooks, model controls and a built Vite app structure.
What I would improve next
Connect real external tools only behind explicit permissions, harden authentication and keep destructive actions confirmation-gated.
Sanitized / excluded
Excluded API keys, Hermes URLs, session secrets, voice provider credentials and any private conversation state.
03

Static mobile prototype with backend-ready paths / Desktop/SHADOW

Shadow

Context
A future-self coaching product built around Mirror, Architect and Commander modes for identity, behavior and daily intervention.
Problem
Most coaching apps collect habits without confronting the identity pattern underneath. Shadow turns behavior into logs, protocols and decisions.
Constraints
Use only verified local evidence, keep secrets out, avoid invented metrics and preserve the project direction already present in the repository.
My role
Conceived the product metaphor, defined the onboarding, directed the terminal-style mobile UI and shaped AI-assisted diagnostic flows.
Product decisions
Built for People who want a direct, structured self-coaching system instead of generic habit tracking. Core features: Diagnostic onboarding, Active directives, Mirror scan, Architect protocol, Commander debugger, Session journal, Evening check-in.
AI workflow
AI-assisted reads and journaling are designed through namespaced Supabase Edge Function calls with local fallback logic when unavailable.
Visual process
Phone-first terminal, doctrine drawer, stark diagnostic language and operating-system metaphors for self-programming.
Technical architecture
Static mobile prototype with local storage, named memory tables and integration placeholders for HealthKit, calendar and screen-time signals.
Iterations
The inspected projects show an iterative pattern: mock first, shape the workflow, add integration placeholders, then verify build behavior before real credentials or customer data enter the system.
Failures or challenges
The safe portfolio version excludes unverified claims, private keys, customer/account information and any feature that exists only as a future idea.
Final solution
A future-self coaching product built around Mirror, Architect and Commander modes for identity, behavior and daily intervention.
Verified outcome
Verified from README and source structure: a working first-version prototype exists with onboarding, local profile storage and AI-uplink design.
What I would improve next
Replace mock nutrition and local doctrine paths with real server-side AI, persistent history and weekly pattern reports.
Sanitized / excluded
Excluded personal memories, book-derived details beyond high-level doctrine labels and any private Supabase configuration.
04

In development / Documents/BOT NT SOFTWARE/PropAccountRotator

Prop Account Rotator

Context
A NinjaTrader add-on architecture for rotating one active prop account at a time with risk controls, dashboards and simulation tooling.
Problem
Prop account management mixes emotional decisions, rule limits and platform friction. The rotator structures account eligibility and risk state.
Constraints
Use only verified local evidence, keep secrets out, avoid invented metrics and preserve the project direction already present in the repository.
My role
Defined product behavior, safety boundaries and workflow architecture while directing AI-assisted C# implementation and review.
Product decisions
Built for Prop-firm traders who need account selection, guardrails and repeatable execution workflows without manually babysitting every rule. Core features: Account discovery, Rotation engine, Risk templates, Execution routers, Bracket management, Dashboard view models, Backtest simulator.
AI workflow
AI-assisted development helped map the domain model, execution boundaries, simulator and product safety language.
Visual process
Operational desktop dashboard rather than a glossy app: account rows, logs, rules and execution state take priority.
Technical architecture
Modular C# add-on with Account, Core, Execution, Risk, Dashboard, Simulator, Storage, Licensing and Strategy layers.
Iterations
The inspected projects show an iterative pattern: mock first, shape the workflow, add integration placeholders, then verify build behavior before real credentials or customer data enter the system.
Failures or challenges
The safe portfolio version excludes unverified claims, private keys, customer/account information and any feature that exists only as a future idea.
Final solution
A NinjaTrader add-on architecture for rotating one active prop account at a time with risk controls, dashboards and simulation tooling.
Verified outcome
Verified from source tree: the project contains a substantial module structure, user guide, QA checklist and release notes.
What I would improve next
Keep the default one-account-at-a-time rotation model, test with simulation/paper accounts and treat any copy-trade mode as separate.
Sanitized / excluded
Excluded trading credentials, account numbers, private performance data, license endpoints and live broker configuration.

Process

Prompt-to-product, not prompt-to-demo.

Across Ink Flow, Shadow, Jarvis and the business tools, the pattern is consistent: turn a loose need into a workflow, direct AI against the workflow, then test where the result breaks.

  1. 01Observe the problem

    Ink Flow began as studio chaos: leads, calls, bookings, payments and follow-up split across tools.

  2. 02Define the user

    Shadow defines the user as someone negotiating with an old identity pattern, not someone who just needs a checklist.

  3. 03Map the workflow

    Prop Account Rotator maps account eligibility, risk state, execution mode and rotation rules before code-level details.

  4. 04Structure the idea

    Jarvis turns the abstract idea of an assistant into lanes, cards, runtime state, speech modes and visible cost.

  5. 05Create prompts and references

    Projects use prompts as product direction: references, constraints, safety rules and expected behavior before implementation.

  6. 06Direct AI-assisted execution

    AI-assisted execution is treated like a directed production crew: useful, fast and still accountable to the product owner.

  7. 07Test real behavior

    Builds and prototypes are checked through local docs, source inspection, build scripts and real UI states where available.

  8. 08Find failure points

    The portfolio excludes private secrets and claims that were not backed by repository evidence.

  9. 09Refine

    Iterations add missing states, fallback logic, mobile behavior and clearer operating boundaries.

  10. 10Deliver the experience

    The finished site packages the work into a clear hiring surface with case studies, lab notes, resume and editing docs.

Supporting Projects

Business tools, visual systems and applied prototypes.

MVP

Hermes / Personal Agent

Telegram-first personal agent with memory notes, web fact-checking, voice/image handling, proactive check-ins and optional NinjaTrader trade alerts.

Node.js, Telegram Bot API, OpenAI, local JSONL memory, MTA data
Prototype

iTrade App

Mobile trading coach app generated from a Figma design, with market pages, AI coach surface, settings and Capacitor notification hooks.

React, TypeScript, Capacitor, Tailwind, local notifications
Prototype

Duality Wellness Web

A visual wellness/cannabis ritual site with strain index, terpene profiles, product storytelling and a polished experiential retail tone.

React, Vite, CSS, product content system
Functional prototype

Pakito Nitro

Automotive business web app with public forms, inventory, repair/sponsor flows and an admin dashboard prepared for Supabase Auth.

Next.js, TypeScript, Tailwind, Supabase-ready auth
Prototype

Cuadrante Facil

Android-first scheduling tool for weekly staff coverage, incidents, change approvals, admin controls and assistant-ready operations.

Flutter, mock data, Firebase/OpenAI-ready architecture
Experiment

Nasty Nash Arcade

Creative music/game environment with Swift tooling and custom audio assets, used as an experimental brand and interaction lab.

Swift, audio assets, local game tooling

Brand + Motion

I build identities that look sharp, move with purpose and stay recognizable.

From the first logo sketch to a complete brandbook, I develop the visual language around a product: aesthetics, color systems, typography, image direction and the rules that keep everything coherent. I also edit professional video and create AI-driven characters and stories for branded content.

01

Brand Development

Positioning, visual territories and identity systems built to make a product feel distinct and coherent.

02

Logo + Aesthetics

Logo concepts, typography, art direction and visual decisions that give the brand a recognizable character.

03

Color + Brandbooks

Color palettes and practical brand guidelines that explain how the identity should look, move and communicate.

04

Photoshop Production

Campaign assets, image treatments, compositing and polished visual production for digital brand systems.

05

Professional Video Editing

Adobe Premiere Pro editing with rhythm, narrative structure, sound, pacing and platform-aware delivery.

06

AI Characters + Stories

Character-led concepts and AI-assisted storytelling shaped into memorable branded video experiences.

LOGO DESIGNBRANDBOOKSCOLOR SYSTEMSPHOTOSHOPADOBE PREMIERE PROAI CHARACTERSSTORY EDITING

AI Lab

Smaller experiments, honest labels.

AI assistants

Jarvis, Hermes and Shadow explore different agent surfaces: tactical command center, Telegram operator and self-coaching system.

What I learned: the interface has to make risk, state and limits visible before AI feels useful.
Voice interfaces

Browser speech, Telegram voice note transcription and optional audio replies appear in prototype form.

What I learned: the interface has to make risk, state and limits visible before AI feels useful.
Trading automation

NinjaTrader alerting, prop account rotation and strategy-lab experiments are present but intentionally not portfolio-dominant.

What I learned: the interface has to make risk, state and limits visible before AI feels useful.
Messaging integrations

Telegram is implemented in Hermes; WhatsApp is treated as unverified/future unless project evidence is added.

What I learned: the interface has to make risk, state and limits visible before AI feels useful.
Dynamic dashboards

Ink Flow, Pakito Nitro and Prop Account Rotator all use dashboard shells for operational clarity.

What I learned: the interface has to make risk, state and limits visible before AI feels useful.
3D environments

Jarvis includes Three.js as an interface layer; separate SketchUp, Twinmotion, Unreal or architecture work was not verified in the scanned repositories.

What I learned: the interface has to make risk, state and limits visible before AI feels useful.
Business tools

Cuadrante Facil, Pakito Nitro and Ink Flow show practical workflow systems for real small-business operations.

What I learned: the interface has to make risk, state and limits visible before AI feels useful.
Market intelligence

iTrade and Personal Agent contain market/news/trading assistant surfaces without exposing private financial data.

What I learned: the interface has to make risk, state and limits visible before AI feels useful.

About

Spanish-born, New York-based founder building with taste, pressure and iteration.

J NASH is a Creative Technologist, AI Product Designer, Brand Developer and Founder with a hospitality-grounded instinct for real-world operations. His work moves between product direction, identity systems, logo and brandbook development, automation, professional video editing and functional prototypes. He uses Photoshop, Adobe Premiere Pro and AI character workflows to turn ideas into coherent brands, cinematic stories and usable experiences.

Contact

Bring a raw idea, a broken workflow or a product that needs a pulse.

New York, NY. The only contact method is email.