Octopus

● The Problem

● The Process

● The System

● The Impact

Email

LinkedIn

Resume

DELHI IST · 00:00:00

Octopus

Octopus

Octopus

YEAR
2023–2025

CATEGORY
PRODUCT DESIGN // INTERNAL TOOLS

ROLE
SOLE PRODUCT DESIGNER

TOOLS
FIGMA // FRAMER

HERO SCREEN
THE ECN DASHBOARD

ID 001

ABOUT

Euler Motors builds electric cargo vehicles in India the three-wheelers and small trucks that move goods around cities. Octopus is the internal system that tracks every one of them, from the first bolt on the assembly line to the last service visit. Over two years I rebuilt it.

Euler Motors builds electric cargo vehicles in India the three-wheelers and small trucks that move goods around cities. Octopus is the internal system that tracks every one of them, from the first bolt on the assembly line to the last service visit. Over two years I rebuilt it.

STORYBOARD, 5 FRAMES

ID 002

A CHANGE APPROVED IN MARCH REACHED THE ASSEMBLY LINE IN MAY.

CONTEXT

[FILL: your real scene] A change to a mounting bracket was approved in a meeting in March. In May a technician was still fitting the old part, because the paper notice never reached his station. Nobody had done anything wrong. There was just no way to see where the paper was.

[FILL: your real scene] A change to a mounting bracket was approved in a meeting in March. In May a technician was still fitting the old part, because the paper notice never reached his station. Nobody had done anything wrong. There was just no way to see where the paper was.

USERS

Three people, one system

Three people, one system

ID 003

Line technician.

Needs to know which version of a part goes on the vehicle in front of him. [FILL: real constraint]

ID 004

R&D engineer.

Needs to trace one component across every vehicle it was fitted to. Used to ask a developer to run the query.

ID 005

Department head.

Approves changes across five teams. [FILL: real constraint]

PROBLEM

Octopus mirrored the database, not the work. Every question an engineer had, a developer had to answer.

Octopus mirrored the database, not the work. Every question an engineer had, a developer had to answer.

Four years of features had been added to a tool built in a hurry. Nothing had been removed.

Four years of features had been added to a tool built in a hurry. Nothing had been removed.

OLD INTERFACE, ANNOTATED
REAL BEFORE-STATE SCREENSHOT WITH 3–4 CALLOUT PINS

ID 006

I was briefed to make it faster and better looking. What it actually needed was for the state of the work to be visible — who changed what, when, and who still hadn’t seen it.

I was briefed to make it faster and better looking. What it actually needed was for the state of the work to be visible — who changed what, when, and who still hadn’t seen it.

PROCESS

Getting a paper process off paper

Getting a paper process off paper

Changing one part sets off a chain reaction. Supply chain sources it, manufacturing retools, quality re-tests, service stocks the new spare. That chain used to run on signed paper moving between desks. I designed the digital version — six stages, each with named approvers and a visible status.

Changing one part sets off a chain reaction. Supply chain sources it, manufacturing retools, quality re-tests, service stocks the new spare. That chain used to run on signed paper moving between desks. I designed the digital version — six stages, each with named approvers and a visible status.

001

Design Change Request

Someone proposes the change

002

Process Change Notice

Departments weigh in

003

Design Release Notice

Cost, time and manpower impact

004

Design Implementation Note

Five department heads sign off

005

Implementation Timelines

When it hits the line

006

Approval Matrix

Final sign-off, up to CEO

DECISIONS

WHAT I HIT

WHAT I DID

WHAT IT COST

Approvals stalled because nobody knew a request was waiting for them.

The dashboard opens on what’s blocked on you, not on everything in the system.

Making pending approvals public created pressure to approve fast. [FILL: how you handled rubber-stamping]

[FILL: second decision — what I hit]

[FILL: second decision — what I did]

[FILL: second decision — what it cost]

[FILL: third decision — what I hit]

[FILL: third decision — what I did]

[FILL: third decision — what it cost]

ECN SCREEN, BEFORE

ID 007

BEFORE

ECN SCREEN, AFTER

ID 008

AFTER

PROTOTYPE

ECN WALKTHROUGH VIDEO
20 SECONDS · MUTED LOOP · POSTER FRAME

ID 009

GLOBAL DASHBOARD

Type a chassis number, get everything that ever happened to that vehicle. This used to live in three spreadsheets and one engineer’s memory.

ID 010

VEHICLE LIFECYCLE

Where every vehicle is: manufacturing, logistics, or after-sales.

ID 011

EULER RELEASES

The compatibility matrix answers software and hardware fit in colour before you read a word.

ID 012

OTA UPDATES

Sending software to a working fleet, with a rollback path.

ID 013

THRUST DESIGN SYSTEM

The component library Euler’s other internal tools now build against.

ID 014

JUDGEMENT

Vehicle testing originally ran inside Octopus in the browser. Browser security limits meant it couldn’t do what the plant needed, and the tools had to be shared with outside vendors. I argued to pull it out and rebuild it as a dedicated Windows app — Merlyn. Results still surface back inside Octopus. [FILL: who pushed back and how you won it]

MERLYN SCREEN

ID 015

IMPACT

What changed

What changed

BEFORE — A change request moved on paper across five departments with no way to see where it was.

AFTER — Every request has a stage and a named owner anyone can check.

BEFORE — Engineers messaged developers to pull data.

AFTER — That data is self-serve.

BEFORE — No record of who had seen a change.

AFTER — Every stage timestamped and attributable.

Octopus is an internal tool with no analytics instrumentation, so I’ve described the change in workflow terms rather than claiming numbers I couldn’t measure. Defining success metrics before launch is the first thing I’d fight for if I did this again.

Octopus is an internal tool with no analytics instrumentation, so I’ve described the change in workflow terms rather than claiming numbers I couldn’t measure. Defining success metrics before launch is the first thing I’d fight for if I did this again.

REFLECTION

What I’d do differently

What I’d do differently

[FILL: a stage you over-designed or a user group you found late.]

[FILL: a stage you over-designed or a user group you found late.]

[FILL: a handoff that broke and what you would change.]

[FILL: a handoff that broke and what you would change.]

Next Project

Next Project

© 2026 ALL RIGHTS RESERVED

MADE IN DELHI, IN