AR.OS1

Advaaith
Ramanan

industrial designgraphic designDAAP · University of Cincinnati

I design functional, engineered products with a focus on problem-solving and system-level thinking. Right now that means one question: what does a physical product become when it's designed AI-native from the start?

More about me ↗

Selected Work

01 · PEA.PODRAISING · OPEN ↗
Pea Pod table pod showing 'Someone's coming to your table'
Tapping Pea Pod's screen to talk
Pea Pod suggesting dishes with prices
Pea Pod checking an allergy question with the kitchen

Pea Pod.

Tables that listen. An AI ordering pod for restaurants, plus its device UI, staff apps, consoles, and design system.

AI-nativevoice + touchUI/UXdesign system
02 · MEERAAN.AIRAISING · OPEN ↗

meeraan.ai

An AI-native family office managing investments, real estate, structuring, governance and concierge.

service designsystems design2026
03 · PLATYPUS.BIRDHOUSEOPEN ↗
Platypus Birdhouse prototype photographed in the studio
Prototype fins, close up
Render with fins open
CAD isometric view

Platypus Birdhouse.

A seasonally adaptive, fully 3D-printed birdhouse for the Northern Cardinal, driven by a 360° rotating fin mechanism.

PETGmechanismFusion 360prototype
04 · GENEVA.LAMPOPEN ↗
Geneva lamp in its arched configuration
Geneva from above
Geneva in an S configuration
USB-C port close up

Geneva.

A modular lamp built from rotating segments that reconfigures into a desk light, a pendant, a wall piece, or a bookend.

lightingmodularbluetoothcompanion app
PROJECT 03PRODUCT DESIGN // INDUSTRIAL

Platypus
Birdhouse.

A seasonally adaptive, fully 3D-printed birdhouse for the Northern Cardinal. A rotating disc opens the fins for summer airflow and closes them for winter.

Platypus Birdhouse prototype, fins open
REFRESH.LOG×
v2 · in progress

The next version reworks the birdhouse's functionality as an AI-native product.

PROBLEM×

Design a seasonally adaptive birdhouse that handles ventilation, predator exclusion, and easy assembly.

01

Research

Line drawing of a Northern Cardinal

Cardinals nest 3 to 10 feet up in dense shrubs and avoid fully enclosed boxes. They want cover but need to see out and leave fast.

Body length 8.3 to 9.4 in, wingspan 9.8 to 12.2 in. Those numbers set the entrance aperture, chamber volume, and fin clearance. Cincinnati sits in the core of their range.

02

Ideation

SKETCHBOOK×
Ideation sketches
03

Disc latch mechanism

TOP VIEW▶
SIDE VIEW▶
ANGLED VIEW▶

Raised knobs on a rotating disc push each fin open as they pass, then release it to fold back against its neighbor. No single position locks it, so it works from any angle.

COMPONENTS×
360° discCarries the contact knobs
Contact knobsTrigger each fin
FinsOpen and collapse in sequence
Flap alignerGuides the collapse
04

Systems

SECTION.ANNOTATED×
Annotated section analysis
DRAINAGE×

Channels built into the top piece shed water outward. No extra parts, nothing pooling inside.

LIGHT REDUCTION×

A 50% antiglare panel at the entrance cuts summer heat and harsh light in the chamber.

HALO LIGHT×

Lights up only at night when there's movement inside. Magnetic mount, one charge lasts a year.

05

Drawings

Annotated render views Annotated plan views

Seven-view projection from the Fusion 360 assembly. Dashed lines reveal the cage, panel inserts, beam channels, and base plate.

LINE.VIEWS×
Orthographic line views
06

Prototyping

Internal hexagonal cage detail
Studio shoot at DAAP

Shot in the DAAP studio. Teal, black, and orange PETG with a waterproof wood-veneer body, held together without glue or fasteners.

07

Final renders

Final render
Fins open, summer
Fins open: summer airflow
Fins closed, winter
Fins closed: sealed for winter
Top view Bottom view
PROJECT 04PRODUCT DESIGN // LIGHTING

Geneva.

A futuristic light that behaves like one. Its rotating chain adapts to the task, the room, or the moment, while staying heavy in the hand and built to last.

Geneva rendered as ASCII art
SPEC.SHEET×
FormIdentical segments on rotating collars
Light2 Cree LEDs per beam, diffused
Power2 × AA, rechargeable or wired
ChargingUSB-C on each beam
ControlRotary dial + Bluetooth controller
AppRooms, scenes, programmable remotes
REFRESH.LOG×
v2 · in progress

Geneva is being redesigned as an AI-native product.

TOP.VIEW×
Geneva from above
01

Ideation

SKETCHBOOK×
Ideation sketches exploring Geneva's segmented form
02

Rotation & reconfiguration

Geneva's segments rotating through a full range of motion

Each joint turns independently, so the same lamp becomes a desk light, a wall-leaner, a hanging pendant, or a bookend.

Geneva front, top and side views
03

Electronics

WIREFRAME×
Wireframe of Geneva's internal electronics
CONTROLLER×

Hanuman smart controller. Bluetooth control of per-side brightness, color, and modes like flash or sound-reactive.

BATTERY×

Double-AA holder built into the base plate. Lifts out in one motion.

ROTARY DIAL×

Pull out to cycle colors; push in and turn to dim.

CREE LEDS×

Two per beam: high output without the bulk.

DIFFUSER×

Softens each LED cluster into an even glow.

USB-C×

One port per beam, so each half charges on its own.

HOOK MOUNT×

Clicks into new positions, shifting the profile and where the light points.

04

Companion app

Group lamps by room, switch them one at a time or all together, apply presets like Sunrise or Neon in one tap, and program physical remotes button by button.

APP.PREVIEW×
Companion app screens: groups, scenes and remotes
05

Final renders

Geneva, arched configuration Geneva, S configuration Geneva, end caps and LEDs
Geneva, USB-C port Geneva, base close up
PROJECT 01PRODUCT DESIGN // AI HARDWARE // UI/UX

Pea Pod.

Tables that listen. A small pod on each table takes orders by voice or touch. It never guesses: when it isn't sure, or a guest mentions an allergy, it holds the order and gets a person.

Pea Pod front view: 'Someone's coming to your table'
STATUS.LOG×
Piloting · raising to scale

60 pods in three of Bengaluru’s leading restaurants, going live in the coming weeks. Hardware, device UI, apps, consoles and launch site designed end to end.

PROBLEM×

Ordering, asking a question, or getting a server's attention all wait on staff being free. Pea Pod gives every table a direct line to the floor and the kitchen.

01

Talk or tap

A guest tapping Pea Pod's screen to talk

Every screen has one obvious action. Tap it, or just say what you want.

Speech recognition tuned for Indian languages and accents turns what a guest says into an order, and a language model reads it the way a server would before it reaches the kitchen.

02

Hardware

Pea Pod rear and front
ESP32-S3×

Runs the touchscreen, audio, and the wireless link to the kitchen.

MIC + SPEAKER×

A microphone behind the front grille and a speaker behind the perforated rear panel.

RGB LED×

Glows through the frosted shell, so the pod reads as on and alive from across the table.

BATTERY×

Runs on battery through service and charges overnight. No cables on the table.

A frosted shell over the screen and an angled base that tilts it up toward the diner. The case is 3D printed.

Pea Pod three-quarter view
Pea Pod held in hand while charging

Pods come off the tables at close and charge overnight, so they're full for the next service.

03

Design language: PEA.OS

A retro desktop: small windows with dark title bars, bitmap labels and 1-bit dithered textures, in greens.

Colour has one job each. The same rules run from the 320×240 device screen to the owner's dashboard. Square corners, no shadows; Schibsted Grotesk and VT323.

COLOUR.ROLES×
LimeNeeds attention, and only that
PeaDone
MintSelected or calm
GreyNeutral
ForestSurfaces and title bars
04

The device screen

23 states, from asleep to allergy hold, on a 320×240 screen.

Big type, flat colour, one moving element per screen, readable at arm's length in a dim room. The allergy hold is calm on purpose: lime, no red, no flashing.

Device screen: listening Device screen: speaking Device screen: allergy hold Device screen: staff on the way

Setup is one flow: scan, pick a table, done.

The pod never asks for a password. The manager scans the QR in the Pea Pod app and picks the table; the phone passes the WiFi to the pod. No other apps, keys or codes to type.

Setup step 1: scan me Setup step 2: phone nearby Setup step 3: joining Setup step 4: ready Setup step 5: couldn't join

If it fails: S3 → S5. One error screen, one fix. The app says which: wrong password, pod can’t hear the phone, or the pod already belongs to another restaurant.

05

Apps and consoles

One app. What you see depends on your role.

Staff phone, light: live alerts and tables Staff phone, dark: live alerts and tables
STAFF PHONE×

"I'll go" in one tap.

Pass tablet
PASS TABLET×

Large enough to read across the pass.

Kitchen tickets with allergy flagged
KITCHEN SCREEN×

Tickets in order, allergies flagged.

OWNER INSIGHTS×

See what guests wanted, not just what they ordered.

Owner insights dashboard
Support console help desk with AI summary and draft
Support console runbook, dark Owner-admin finance console
SUPPORT CONSOLE×

A live queue, step-by-step runbooks, and AI that only drafts. Nothing sends until a person presses Send.

06

Rules I designed to

RULE_01×

As few taps as possible.

RULE_02×

One main action per screen.

RULE_03×

Plain words.

RULE_04×

Errors that say what to do.

RULE_05 · SAFETY×

Safety steps stay deliberate: an allergy hold always needs a named person and two taps.

07

Try it live

The real prototypes, running in the page. Every animation is live. Switch programs below, or open one full screen.

PEA.OS 1.0 · SCREEN.EXE× ·
Every device screen state at real size, animated, plus the QR setup flow.Open full screen ↗
08

From simulation to service

A restaurant at 9 pm on a Saturday is the only real test. So Pea Pod learned in rehearsal, and is now learning on the floor.

Before it met a guest, Pea Pod was trained in a near-realistic simulated service: noisy rooms, half-finished sentences, allergy questions, changed minds. Now it runs in real dining rooms, because only real service produces data worth learning from.

ROLLOUT.LOG×
01SimulateTrained in a near-realistic restaurant environment before touching a real table.
02Pilot · now60 pods in three of Bengaluru’s leading restaurants, self-funded.
03Go liveFull service in the coming weeks.
04Scale50 restaurants and 2,000 pods across Bengaluru.
3
restaurants piloting
60
pods built, self-funded
50
restaurants next, in Bengaluru
2,000
pods to install
09

We’re raising

Sixty pods proved we can build it. The next round puts 2,000 on tables across 50 Bengaluru restaurants.

TRACTION×

Trained in simulation, now piloting in three of the city’s leading restaurants and going live within weeks. Built and funded to this point on personal capital.

MODEL×

One fee, once, covering hardware and installation. Restaurants say yes without a long contract, and every pod pays for itself up front.

THE MOAT×

Decision data, not recordings: what guests ask, what they choose, when they need a person. Every service makes each pod better at its room and its regulars.

THE ASK

Capital to scale from 60 pods to 2,000.

Talk to us about investing →
10

Final renders

Pea Pod, pink accent Two Pea Pods, lime accent: allergy check and charging
Pea Pod on a table Pea Pod in hand
PROJECT 02SERVICE AND SYSTEMS DESIGN · 2026

meeraan.ai

An AI-native multi-family office that delivers the full spectrum of family office services through a coordinated team of intelligent agents.

Twenty years of family office operations, encoded into AI agents. meeraan.ai gives every family the rigour of a full institutional team, without building one.

MY ROLE×
Designer and director

I defined how the service should behave, set the rules it runs on, designed the family app and the office console, and led the build and testing.

01

What it does

INVESTMENTS×
Investment strategy and advisory
Asset allocation
Manager research and operational due diligence
Real estate investment and property management
WEALTH + REPORTING×
Wealth structuring
Legal and tax coordination
Consolidated performance reporting
Family accounting
FAMILY×
Family and corporate governance
Succession planning
Next-generation education
Philanthropic advisory
EVERYDAY×
Concierge and lifestyle management

Institutional-grade oversight of a family’s entire wealth ecosystem, delivered with continuity, transparency and discretion, at a fraction of the cost of a traditional family office.

02

How it learns

20+ years of family office decision-making experience trained the rule-making layer around the model.

Two decades of decisions, negotiations, processes and approvals shaped its rules, memory, specialists and checks. The model underneath does the reasoning. A person still makes the call.

FOUNDATION MODELS: CLAUDE (ANTHROPIC) · CODEX (OPENAI) · MODULAR, OPEN-WEIGHT READY

03

What guided the design

PEOPLE DECIDE×

People decide.

The system prepares; a person approves.

PRIVACY×

Privacy by default.

Each family member sees only their own world.

LANGUAGE×

Plain language.

No jargon, no system terms, ever.

NEXT STEP×

Always a next step.

Every request ends with something clear: work ready to review, one question, or a person who has it.

CALM×

Calm by design.

A quiet visual language: warm paper, black ink, one accent colour, because money deserves calm.

FIVE SECONDS×

The five-second test.

Anyone looking at any screen should know where they are, what’s happening and what to do next.

04

We’re raising

Some of the world’s largest private fortunes are still run on spreadsheets, inboxes and a few senior heads. We’re building the agentic family office.

OPPORTUNITY×

Agentic family offices are a category almost no one is building for. Families want the ease of “just ask”; offices need it to stay strict, private and accountable.

THE EDGE×

Two decades of real decisions, negotiations, processes and approvals. A modular layer that runs on today’s best models and isn’t locked to any one.

STAGE×

A working office in private development: family app, office console and a team of specialist agents, with a person approving every decision.

THE ASK

Raising seed capital to take meeraan.ai from private development to its first 100 clients.

Talk to us about investing →
05

What I learned

01

The interface is the last 5%.

Most of the experience is decided in places nobody sees.

02

Rules are design material.

Writing them down early made every later decision easier.

03

Test as the user, not the maker.

My most useful findings came from using it the way a confused family member would.

04

Fix the system, not the symptom.

Every failure points to a path that will fail again.

NDA×

meeraan.ai is in private development. Details, including the data room, are shared with investors under NDA.

GRAPHIC DESIGN03 POSTERS

Posters.

POSTER_01 · COCKATOO×
Cockatoo instructions poster
POSTER_02 · FUTURE×
Future poster
POSTER_03 · OGRAL×
OGRAL poster
CINCINNATI, OH+1 (513) 266-8837

Resume.

↓ Download PDF
EDUCATION×
University of Cincinnati (DAAP)
B.S. Industrial Design · 2024 to 2030
  • Top-ranked industrial design program
  • International Outreach & Global Scholarships
  • Co-op: alternating semesters of study and industry work
Rhode Island School of Design
Pre-College Summer Program · Jun to Aug 2023
  • Competitive industrial design program
  • Concepts in Procreate, Photoshop, and Shapr3D
  • Mentored by senior designer Paul Osimo
Stonehill International School
Bangalore, India · Class of 2024
EXPERIENCE×
Industrial Design Intern, Foley Designs
Bangalore, India · Jun to Aug 2025
  • Early-stage concepts and research for a portable industrial battery system
  • Sketch concepts and visual narratives for product direction
  • CMF strategies and prototyping approaches
  • Multidisciplinary sprints across consumer tech and mobility
CERTIFICATIONS×
AI/ML for Autonomous Drone Systems
Accelerated Professional Program · Completed
computer visionSLAMsensor fusionpath planningreinforcement learningdrone swarms
DGCA Certified Remote Pilot, Quadcopter
Directorate General of Civil Aviation · India
TOOLS×
Software
Fusion 360Shapr3DSketchUpVizcomProcreatePhotoshopIllustratorInDesignOrca Slicer
3D printing
SLAFDMVAT photopolymerization
AWARDS + ACTIVITIES×
  • UC Global Scholarship, 2024
  • International Outreach Scholarship, 2024
  • High School Design Competition, 2nd Place, 2024
  • Varsity Soccer and Varsity Track, 2023 to 2024
INDUSTRIAL DESIGNERCINCINNATI, OH

About.

ADVAAITH.RAMANAN×
Portrait of Advaaith Ramanan
Fun fact
DGCA-certified remote pilot (quadcopter).
Fun fact
Varsity soccer and varsity track, 2023 to 2024.

I'm studying industrial design at DAAP, University of Cincinnati. My focus is functionality and problem-solving: products that are practical, efficient, and thoughtfully engineered.

With an engineering background, I'm drawn to complex products that need technical understanding, precision, and system-level thinking.

I have a keen interest in developing AI-native products, where the hardware, sensing, and software are designed together from the start.

Function comes first, but I keep a strong balance between form and function so the result feels professional and considered.

EXPERTISE.SYS×
industrial designgraphic designAI-native productsmechanisms3D printingCMFprototyping
LinkedIn ↗ Instagram ↗ Resume
Copied to clipboard