Thomas O’Donahoo and the art of thinking small

Thomas O’Donahoo once filled in for a podcast host and immediately encountered a problem of scale. The regular host, Simon, was six foot five. Filling his shoes, O’Donahoo explained, would be difficult in both senses. It was a small joke before a serious conversation with teacher Yasodai Selvakumaran, and a useful introduction to someone whose business involves making large things manageable. A whole syllabus. A classroom’s worth of different starting points. The distance between hearing an explanation and actually understanding it.

At Atomi, the education company he co-founded with Robert Barakat, O’Donahoo has spent years working on that distance. The business began with short explanations of mathematics and grew into a platform for high-school lessons, practice and assessment. His public work moves between software, teaching and the question of what a student should do next. The examination result remains part of the picture. He is especially interested in everything that happens before it.

Two friends, one lesson at a time

O’Donahoo and Barakat were friends at St Aloysius’ College in Sydney, both members of the class of 2009. Their conversations about business began at school. Education was absent from those early ideas. There is something pleasantly ordinary about that detail: the eventual occupation did not arrive neatly labelled during adolescence. It emerged through work the pair could already see around them.

After school, Barakat organised a tutoring business with fellow former students. Holiday sessions for Year 12 pupils gave the friends a practical starting point. They began considering how video could take explanations beyond the limits of a tutoring appointment. A recording could travel further than the person delivering it, and a student could return to the difficult bit without booking another session.

Atomi was founded in 2013, while the pair were at university. The early maths videos broke complicated ideas into smaller pieces. That choice would remain recognisable as the business expanded. Years later, its catalogue would reach far beyond those first explanations, but the original unit of work was modest: take something confusing and explain it clearly enough that somebody could use it.

A name with small ambitions

In 2016, O’Donahoo explained the thinking behind the Atomi name. He reached back to the idea of the atom: break a complex thing into its fundamental parts, then understand how those parts fit together. Applied to education, the metaphor describes a method rather than a decoration. A learner needs the pieces before the elaborate structure can make sense.

“We create atoms of knowledge.”Thomas O’Donahoo, 2016

His examples ranged from music to the physical world. The point was that complexity becomes less forbidding when you can recognise what it is made of. The company’s name puts that conviction on the door. It also sets a demanding brief. Calling a lesson short is easy. Deciding what belongs in it requires a view of what the student must understand first.

There is an editorial judgement inside every explanation: what to introduce, what to postpone, which example earns its place. O’Donahoo’s description of Atomi makes those choices central to the enterprise. A video library can accumulate endlessly. A useful learning sequence has to establish relationships between its parts. The atom metaphor offers a way to talk about that work without requiring a learner to admire the machinery.

Carbon, Nucleus and the work underneath

O’Donahoo studied commerce and science at the University of Sydney, with finance, financial mathematics, statistics and computer science among his majors. The combination places commercial questions beside technical ones. In his published work, he has been willing to get quite specific about the latter. In May 2017, he opened up Atomi’s software stack for readers curious about how the platform worked.

The web client was called Carbon, the backend Nucleus, and the marketing website Hydrogen. Someone had evidently allowed the naming theme into the engineering department. The inventory included JavaScript, a PHP backend using Laravel, MySQL and a range of services for payments, video and operations. It was a snapshot of the company at that time, with the everyday dependencies left visible.

He presented the list partly to give credit to the products and teams Atomi relied on. That is an appealingly concrete form of acknowledgement. A lesson reaching a student involves far more than pressing play: accounts, delivery, records, subscriptions, testing. His account treated these as parts of a working system. The naming was playful; the list underneath was businesslike.

For a profile of a founder, this is useful evidence of where he has directed his attention. Alongside writing about learning, O’Donahoo has explained the infrastructure behind it. The two concerns meet whenever a student expects a lesson to be available and a teacher expects the accompanying information to be useful. Reliability is a quiet part of the educational promise.

When the teacher becomes the customer

By 2017, Atomi’s maths beginnings had become a wider HSC resource business. O’Donahoo and Barakat were included together in The Educator’s Hot List that year. Its account described more than a million lessons delivered to more than 60,000 students, with videos used by over 5,000 teachers across New South Wales. Those are historical figures, a view of the platform during its earlier expansion.

The same year, the founders brought in senior reviewers, including experienced teachers, curriculum writers and HSC markers, to work with content producers. The arrangement matters because an explanation has several audiences. Students need it to make sense. Teachers need to trust its relationship to the curriculum. Someone has to consider both, preferably before the lesson reaches the classroom.

O’Donahoo’s product announcements show the classroom becoming a more explicit design problem. In August 2020, Tasks 2.0 let teachers combine different Atomi items with their own links and attachments. Work could go to a whole class, a group or an individual. The details sound unglamorous until you imagine preparing different work for different pupils. The small controls are where an intention becomes something a teacher can actually do.

The September 2022 update added drafts, scheduling and duplication across classes or teachers. A draft is hardly a stirring technological prophecy. It is, however, a concession to the fact that people get interrupted. Reusing an assignment acknowledges that useful preparation should survive its first outing. O’Donahoo’s announcements describe a platform learning to accommodate the routines of the adults who teach with it.

The money had homework attached

In December 2021, Atomi announced an A$9 million Series A led by OIF Ventures. The stated uses included more content, more year groups and further platform development. Funding entered the story with a list of jobs to do. The company was already serving students and teachers in Australia and the UK; the investment was intended to support further work in those markets.

OIF described the founders as complementary, emphasising Barakat’s entrepreneurial approach and O’Donahoo’s technical ability. It also highlighted their interest in research-based resource design and information teachers could act on. That description fits the published work: an explanation on one side, a system for understanding the learner’s progress on the other. The founders had to build the connection between them.

Relationships with educators extended beyond individual customers. O’Donahoo announced a partnership with the NSW Secondary Principals’ Council in November 2021, aimed at creating opportunities for principals to connect, share practice and advocate for schools. A learning platform enters an institution with its own responsibilities and habits. The partnership placed Atomi alongside a professional community already doing that work.

Feedback while the answer is still warm

O’Donahoo’s public comments about assessment distinguish the final examination from the process of learning. He recognises the value of a common exam as a measurement of attainment. His interest is in giving students useful responses throughout their work, while teachers can see what is happening. A mark alone leaves the next move unspecified. Feedback needs to suggest something the learner can act on.

That ambition explains his interest in AI more precisely than a general enthusiasm for new technology would. The desired sequence includes marking, personalised guidance, an opportunity to use it, and teacher oversight. Each step has a purpose. Speed helps when it shortens the interval between an attempt and a useful correction. The objective remains the student’s next piece of work.

In an October 2024 article about AI cheating, O’Donahoo argued for addressing the incentives around students’ behaviour: pressure to achieve, tasks that feel too difficult or dull, and classrooms where admitting uncertainty is uncomfortable. His advice included honest conversations, engaging work and feedback. The approach asks what would make a genuine attempt worthwhile to the student. It takes the question back into lesson design.

For the 2025 National Education Summit in Melbourne, he was advertised to speak about patterns in learning drawn from more than a million Atomi students. The proposed session concerned engagement, retention and practical changes in teaching. The scale had grown considerably from a handful of maths explanations. The subject still came down to what students understand, and what someone can do to help them understand the next thing.

There is a continuity here that survives the funding figures and product releases. O’Donahoo keeps returning to the parts: an explanation, an attempt, a response, another chance. The business has acquired more machinery to connect them. His original metaphor remains a useful test of the result. Take the big idea apart. Give the learner something they can grasp. Then see what they can build with it.