Building Zero

A year and a half designing a product, a team, and a way to practise a job before you have one.

I joined Zero as the third person in the company, responsible for designing the product from scratch.

Over the next year and a half, we went from a blank Figma canvas to a product about to reach its first users, a global waitlist of 10,000 people, and a brand with more than 20 international recognitions. Along the way, I built a 12-person Design & Experience team.

The brief was to help people learn the skills they would actually need at work. Following that brief took us through simulated companies, AI managers, stakeholder presentations, a small film production team and, eventually, a group of product designers trying to understand why people watched our ads and left before signing up.

Head of Design & Experience · Product, brand, team building, storytelling and growth · Product preparing for its first users at the time of writing.

Start inside the job

The starting point was a gap we kept coming back to. Someone could finish university, arrive at their first job and still need months of training before they could do what the job required. We wanted to bring some of that experience forward.

Design gave me a useful reference. A portfolio lets you show somebody how you think through the work itself. I wanted people entering other professions to have something similarly substantial to bring into a hiring conversation.

That meant working backwards from a role. If someone wanted to become a data scientist, what would they need to use, understand, decide and explain? Where could they practise those things before an employer expected them to know?

In our first version, the learner chose a role and was assigned a simulated company. They entered as an intern, met a manager and worked through situations using the tools they would need on the job. Across the versions, the experience took shape around six parts:

The familiar learning dashboard was tempting. It would have given us a sensible place for modules, progress bars and the next thing to click. I was worried about how quickly that would make Zero feel like another course.

We spent a lot of time trying to give the learner a sense of arriving somewhere. The floating world was part of that early direction, and we held onto it through the first version. A company and a manager gave the task context before the learner started working.

A company became a city

The product kept changing as the backend evolved towards personalised AI learning pathways. V2 had to accommodate more ways for someone to build the skills their chosen role required.

The separate company worlds grew into Zero World, a city containing multiple companies. Moving through that world would take the learner through different projects, tools and concepts. We wanted the ambition of progressing through a game: there was more of the world to take on as you became more capable.

That gave us a much larger design problem. A city could make the journey exciting, but the learner still needed to understand what they were there to do. The connection between a company, its projects and the skills being practised had to remain clear.

We also wanted the companies to feel inhabited. We borrowed from the way games introduce a mission. You arrive at an office, meet the manager and walk into a little of the life already happening there. There might be something funny going on in the background. The characters returned around projects to introduce the work or respond to how it went.

Those scenes gave us room to build a relationship with the manager over time. They also made writing, illustration and motion part of the learning experience from very early on.

We began prototyping behavioural mini-games for the less tidy parts of a job. A learner had to make a decision in a social situation at work. Technical skill would only take them so far if they could not handle the people around the task.

Much of this was still our best hypothesis. V2 grew out of our assumptions about learning and the possibilities opening up in the technology. We were about to find out how those decisions held up with users.

The effort we decided to keep

I initially struggled with the decision to make learners take notes.

The tools around us were making notes disappear. An AI could listen to a meeting and produce a summary while you carried on with your day. My instinct was to make Zero similarly convenient.

The founders wanted learners to take their own notes.

I came around because of what taking a note actually asks you to do. You have to listen, work out what matters and put it into words you can use later. For somebody still learning how to work, that effort had a purpose. Removing it could mean removing part of the practice.

That became a useful complication in the rest of the product. An AI coworker could stay beside the learner while they used a tool or worked through a concept. But if it did too much, the learner could finish with an answer and very little idea of how they got there.

We introduced Proof of Understanding and Proof of Progress to make learners demonstrate what they were picking up as they went. Getting through the experience had to involve applying the concepts and showing that something had been understood.

The same thinking carried into the presentation at the end of a scenario. You had done the work. Now you had to make a case for it to stakeholders.

They could question your choices, push on the reasoning and put you under pressure. You had to explain yourself, take a stand and handle being grilled. We wanted learners to encounter that feeling here, while there was room to practise, before meeting it in an interview or at work.

You can spend a long time learning something privately and still feel completely unprepared when another person starts asking questions. We wanted that part of the job to have a place in the learning too.

Together, the work and those conversations could create a much richer portfolio. It could show the tools someone had used, the concepts they had applied, what they produced and how they handled the discussion around it. The longer-term ambition was to use that evidence to match people to job descriptions and give hiring managers a clearer picture of their capability.

Whether it would be enough to earn that trust was still ahead of us. But it gave each project a reason to end with evidence someone else could inspect.

Finding a face and a voice

Alongside the product, we were figuring out what Zero should look like.

I started with a small team spanning product design, motion and illustration. That mix mattered. We were trying to make an unfamiliar experience feel coherent, and a screen often depended on a character, a piece of motion or a story being worked out beside it.

The logo took longer than I expected.

I had gone into branding thinking mostly about the people outside the company. Would they understand it? Would they remember it? After enough iterations, I began to think that the harder audience was the one inside the company.

The founder had to recognise their idea in it. The rest of us had to feel we could build from it. You can explain a mark to an audience over time; getting people in the same room to agree on what it means can take a while. My job became finding a balance between taste, the founder's point of view and a visual language we could keep using.

Voice onboarding gave that identity another job. We wanted a short conversation to help Zero understand the learner and make the introduction feel personal. We explored orbs, circles, typography and a moving paintbrush, looking for something with its own character.

One direction kept pulling me back. When you spoke to Zero, the interface rippled. When Zero spoke back, a soft inner glow appeared. Both behaviours came from the logo.

For a while, I could only show people an After Effects animation of the experience. About a year later, I rebuilt it as a working prototype with live voice and responsive visuals. The tools had caught up enough for me to try the conversation we had been imagining.

That prototype was still an exploration. Some of the work I care about most from Zero sits in that category.

Thirty minutes was too long

By the time V1 existed, explaining Zero in a room took me around half an hour.

We needed a product video for investors. A screen recording was the obvious starting point, but I wanted to see whether we could explain the experience through a story. I priced out actors, a set and a crew. We could not afford it.

Around the same time, I came across AI videos that held together as narratives. I traced how they were being made and spoke to people working on micro-dramas at Pocket FM. They were organising production into small pods, bringing writing, editing and coordination together. That gave me an operating model I could try.

We assembled our own team. Saba directed, Tushar worked through the generations, Aarav edited, and Sanjay, Rashi and Roshan built pieces of the film.

Character consistency and lip sync made the process difficult. It took hundreds of generations to get the direction right. The ability to generate a shot quickly did very little for us when the next shot looked like it belonged to a different film.

Eventually, we had a four-minute story that could explain Zero.

The models have moved so quickly that parts of it already look dated to me. I still care about that film. We had found a way to make something our budget would previously have ruled out, and the team had learnt a new way to work together while doing it.

The same effort to make Zero understandable ran through the investor decks, memos and demos. Each format forced us to decide what somebody needed to understand first, and what could wait.

Letting people see the work

We were making enough material to keep ourselves very busy and most of it was staying inside the company.

I helped get Zero's social presence going, working with marketing collaborators on AI films, carousels and founder videos. The founder videos gave us space to explain the thinking behind the company. The more experimental films let us try different ways of getting people interested.

Inside the team, we started a build-in-public habit: one designer shared work each day. Product studies, motion tests and unfinished ideas began finding an audience.

That had a side effect I valued. Zero became more visible in the design community as a place where people could make ambitious work. The way we were building was becoming part of the reason people wanted to work with us.

Then we needed people to sign up

When V2 needed beta users, the job changed again. This time, there was no marketing team we could bring in to take over. A group of product designers had to figure out acquisition.

For a while, I was in Vietnam trying to find a video editor. Then I discovered that the people I was already working with could edit. Akshay had learnt filmmaking before design. Rashi had years of motion experience. I had spent months beside this team and still managed to go looking elsewhere for skills they already had.

We made a pigeon ad. Akshay and Rashi pulled it together in two days.

There was plenty of less enjoyable learning too. We tried Meta and TikTok ads, AI UGC, clipping and tools that promised to make parts of the process easier. We got blocked on Reddit. We learnt how much could go wrong between a person seeing a video and joining a waitlist.

Some videos attracted views and produced very few signups. Others got people to click, only for the landing page to lose them. We had to understand the whole route, including what it cost to get someone through it.

Our approach to the videos changed. We began separating the hook, the middle and the call to action so we could change one part at a time. We kept characters consistent across variations, which meant we could combine sections without the edit falling apart.

A weak opening needed different work from a confusing landing page. Once we could see where people were dropping out, we could choose what to change next. Learning to tell those problems apart was a large part of learning marketing.

Through all of that, we built a global waitlist of 10,000 people.

Why Zero

There was also a larger argument we wanted the brand to make.

Putting Zero beside established education platforms and listing the features would leave too much unexplained. We wanted people to understand why we were building this in the first place. That became Why Zero.

The team developed the concept, storyboards and visual direction in-house. Our internal developers could not take on the build, so we found an external partner, BUNQ LABS. Their team helped turn the visual material into an interactive experience.

You drew a zero to enter. Further in, you held to shatter the glass as the familiar promise of education broke apart. The story eventually opened into a city you could explore. The interactions asked the visitor to participate in the argument. The development team's account of the build goes into how they brought those moments to the browser.

The website went on to receive more than 20 international recognitions across awards, nominations and features.

For the team, there was something lovely about seeing it appear in the places we used to visit for inspiration. We had spent so much time looking at other people's work there. Now ours was part of it.

Just before the first users

By the end of this chapter, the work had travelled with me from India to Dubai to Vietnam and back to Dubai. I had spent about a year and a half building Zero and its design team, and I wanted to take a breath and do something new.

I am proud of how much we made, and of the people who made it with me. I had joined as the third person, with a product to design from scratch. I was leaving a team whose work ranged from the smallest interaction to a fictional company someone could learn inside.

There are encouraging signals already: people have signed up, the brand has found an audience, and the work has travelled. The product is now about to reach its first users. Its promise still has to hold up when someone is struggling with a task, asking for help or trying to explain a decision.

I would like to see what happens when one of those learners reaches a real stakeholder meeting. Whether, somewhere in the questions, they recognise something they have practised before.