01 / 06Equal groups
Count the equal groups.
Three groups. Two in each. How many altogether?
See the relationship
6 objectsThe child connects the total to three equal groups, rather than treating the dots as an unrelated collection.
Founder & operatorKen, for short.
Keerthikan Thirukkumar
I work on problems that matter to people, especially when the way forward isn’t clear.
Designed and built Toku.Founded Akura.
I can turn your ambiguous problems into solutions that work.
Measurable impact on people.
How I work
I work out what needs solving, test the assumptions and build. When the evidence changes, so does the approach.
Toku began with children I taught. I designed and built it myself. Other founders started asking for advice, and that became Akura.
I’m driven by work that measurably improves people’s lives. What matters is who benefits and what changes for them.
What that can look like
That could mean learning with more confidence, getting the right care sooner, or spending less time and money on something you need.
Toku and Akura are both part-time.
I designed and built it
Toku
Practice that fits the child.I built Toku to make personalised learning affordable. Children learn the skills behind the questions, then practise using them.
Time to make the questions.
By hand · one question30 min
Toku · one questionOn demandGenerated in the user’s browser.
Fresh questions on demand.
Generated in the user’s browser, without writing each question by hand.
I spent 15–30 minutes writing each question. Compare the time for a set of 20.
For the childThe next question
The next question
can fit the gap.
Practice can target the skill and level the child needs, with new questions ready on demand. They don’t have to wait for someone to find or write another paper.
Parents and tutors spend less time finding and marking materials. They can use that time to explain what the child missed.
About the comparison
The 15–30 minutes is my question-writing time, not an industry average or the child’s practice time. The graph calculates the time for 20 questions.
Toku generates questions in the user’s browser. “On demand” describes availability; the display does not claim a measured zero generation time.
The sheet comparison uses 20 questions at the selected number per sheet, rounded up. Toku offers digital practice and printable sheets. The comparison is an illustration, not a measured paper or carbon saving.
Beyond the answerUse the skill
Use the skill
when the question changes.
These six steps keep 3 × 2, but change what the child has to do with it. They move from counting groups to recognising and using the relationship.
02 / 06A different representation
Find the groups in the rows.
The same six objects are arranged in rows. What stayed the same?
See the relationship
3 rows of 2 still make 6The layout changes. The child has to recognise that three rows of two are still three equal groups.
03 / 06From objects to symbols
Write what the picture means.
How does 2 + 2 + 2 become 3 × 2?
See the relationship
3 groups × 2 in each = 6The child connects the symbols to the number of groups, the size of each group and the total.
04 / 06Inverse reasoning
Find the missing amount.
Three equal groups make six. How many are in each group?
See the relationship
2 in each groupThe total is given. The child uses the relationship backwards to find what belongs in each group.
05 / 06Transfer to a new context
Use the idea in a new setting.
One ribbon is 2 m long. Another is three times as long. How long is it?
See the relationship
6 mThere is no multiplication sign telling the child what to do. They have to recognise the relationship in the ribbon lengths.
06 / 06Generalisation
Describe it without a fixed number.
Each group contains a. There are three groups. What is the total?
See the relationship
3a = a + a + aThe group size is now a letter. The child has to explain why three equal groups add up to 3a.
01 / 06 · Equal groups
Swipe the examples or choose a step. Watch how the reasoning changes.
The same idea. A new way to use it.
I designed the generation rules to change how an idea is presented and what the child has to work out. A child can practise with support, then use the same skill in a less familiar question.
This six-step example shows the approach. It is a demonstration, not the full learning engine.
Parents kept asking for better practice.
The materials they could buy were often expensive or wrong for their child. The topic might fit, but the level, format or length didn’t.
I made personalised materials and saw children improve. But writing one question took me 15–30 minutes. I could keep the price down by putting in my own time. I couldn’t reach many more children that way.
I designed and built Toku myself.
I built the learning experience, the platform and the question-generation system. There was no team behind me. The business started as Exam Bytes Academy and is now Toku.
I wanted practice to address the skill a child was missing. Changing the numbers wasn’t enough. The questions had to change how the child used the idea, so remembering an answer would only get them so far.
I kept tutoring. It paid for the build and kept me close.
Teaching funded development. It also meant I was still working with children and talking to parents while I built.
I could see where attention dropped, which explanations helped and when a child got stuck because a familiar question looked different. Those lessons gave me things to test in the product.
It started with exam access. Confidence mattered too.
I wanted children to get good preparation without the cost shutting them out. But I’d also seen children decide that maths, English or reasoning wasn’t for them before they’d had a fair chance to learn.
Toku grew beyond exam practice. I wanted children to understand the basics, use them in unfamiliar questions and see that they were capable of learning more.
The steps I changedI cut the work
I cut the work
around the practice.
Before a child starts, someone has to find the right paper. Afterwards, someone has to mark it and work out what comes next. I built Toku to shorten those steps.
A printed-paper route →︎ Toku on screenSkip comparison ↓︎
The question is ready when the child is.
A printed-paper route
Someone writes the questions, assembles the paper, then prints and distributes it.
With Toku
Toku generates questions in the user’s browser for the skill and level being practised.
Each new question no longer needs writing by hand.
Start practising when it suits you.
A printed-paper route
A parent finds a paper, buys it and brings it home, or waits for delivery.
With Toku
A parent signs up and adds their child. Practice is ready on screen. Printable sheets are available too.
Digital practice has no delivery wait.
Practice matched to the skill and level.
A printed-paper route
An adult has to check whether the paper fits the child’s needs. The year group alone may not tell them.
With Toku
Checks find a starting point. Practice targets a skill and level. Guided helps choose the next sprint.
The practice can target the gap.
Make starting feel manageable.
A printed-paper route
A long paper can put the child off before they answer the first question.
With Toku
Short sprints give the work a clear finish. Feedback and rewards encourage the child to come back.
I designed the session around the child using it.
Spend the time explaining, not marking.
A printed-paper route
A parent or tutor marks the paper, finds the errors and works out what needs explaining.
With Toku
Toku records the answers and handles routine marking. Parents and tutors can review the attempt when it suits them.
The attempt is there when the adult needs it.
Use the result to plan the next lesson.
A printed-paper route
Someone finds another paper, checks the fit and sets more work.
With Toku
Parents and tutors use the results to plan support or homework, with more targeted practice ready to generate.
The next step can start with what the child just showed us.
Scroll down to follow the next step
What changes
for each person.
Learn it well enough to use it elsewhere.
Getting the answer once is a start.
A child practises in short sprints, gets feedback and meets the idea in different forms. I want them to understand why it works and recognise when to use it, even when the next question looks different.
See where your child needs support.
See the work, the results and where to help.
Parents can review attempts and see what needs attention. Guided helps choose the next practice; Practice leaves the choice with the adult. Children can work on screen or use printable sheets.
Know what to work on before the next lesson.
Spend less time making and marking homework.
Tutors can set practice, review attempts and use the gaps to plan their next explanation. They can target one skill without repeating a whole topic. The tutor still makes the teaching decisions.
Set practice that fits each child.
Each child can practise at the skill and level they need.
Tutors share the resource, not a single worksheet for everyone. They can set and review homework while Toku generates the questions. Their time goes into teaching the child.
Personalised practice for more children.
I built Toku around keeping personalised practice affordable.
I define the skills, relationships and question rules. The platform uses them to generate practice. Supporting another child or a more specific gap no longer means writing another set by hand.
How Akura started
Akura
I built Toku from scratch. Other founders started asking how I’d handled the problems they were facing. I set up Akura to make that work a regular part of what I do.
Building on the founder’s expertise.
I recognised the questions. What should we build first? Who needs it? How do we turn interest into customers? I’d had to answer them with my own time and money on the line.
I had to make the time and money count.
Tutoring paid for Toku’s development. Every feature took time I could have spent teaching or building something else. I had to choose what would help children most and find a way to keep the practice affordable.
Read all three stages
I had to make the time and money count.
Tutoring paid for Toku’s development. Every feature took time I could have spent teaching or building something else. I had to choose what would help children most and find a way to keep the practice affordable.
Other founders started asking.
They brought problems from their own businesses. I asked questions, challenged assumptions and helped work out what to do next. Sometimes the product needed to be clearer. Sometimes the issue was finding customers.
I set up Akura.
The advice led to more work together. Akura gave it a structure: a clear role, an agreed contribution and an ongoing relationship with the founder.
I work for
equity.
With selected founders, I take equity rather than an advisory fee. I stay involved in the business we’re building.
The work can be product decisions, positioning, customer outreach or an investor pitch. I start with what’s holding the company back.
Selected workWhat I
What I
worked on.
At Credo, I worked on the product and how to explain it. At Wavionix, I helped with customer outreach. For a fashion-sustainability venture, I changed the approach to customer acquisition.
Credo / product and positioning
A clear offer for recruiters.
Credo’s founder knew explainable AI: technology that could show why a hiring recommendation had been made. I worked on turning that capability into a clear offer for recruiters.
I helped define the offer, positioning, pitch, deck and outreach. We separated the first product from the wider ambition for a hiring portal, so the initial pitch had something specific to sell.
Wavionix / customer outreach
Making the case to customers.
Wavionix brought safety-technology expertise from aviation. I helped the founder with outreach to potential clients.
I worked on explaining the offer from the customer’s side: what it would help them do, why they would need it and why it was worth a conversation.
Fashion sustainability / customer acquisition
30×daily new customers
in seven days
in seven days
A different route to customers.
I redesigned customer acquisition for a fashion-sustainability venture. Daily new customers increased 30× in seven days.
Measurement
This measures daily new customers over a seven-day period after the acquisition change, not revenue or a longer-term growth rate.
Combining their expertise
with what I’d learned.
At Toku, I could make a decision and build it myself. At Akura, I work with founders who know things I don’t. I need to understand their customers, their constraints and why they’ve made the choices they have.
I learned to listen before making a recommendation, challenge decisions with reasons and make the next step specific enough to act on. The founder’s expertise matters. So do the time and money available to try something different.
Akura’s website
Short videos on the decisions behind starting a company are coming to Instagram and YouTube.
Instagram @akura_vcComing soon ↗︎YouTube @akura_vcComing soon ↗︎Akura is part-time.Next: difficult decisions →︎
Difficult decisions
Decisions people
can trust.
The parent paid. The child did the work. I was responsible for the result and the business. I had to understand each perspective, explain the trade-offs and keep us focused on the same goal: the child making progress.
The shared goal was progress. My recommendation had to serve it.
01 / 03Making the lesson
Making the lesson
fit the learner.
A parent wanted two uninterrupted hours of tutoring. I recommended shorter stretches with breaks. We both wanted the child to improve. We disagreed on how.
The request and the decision.
We shared the goal. The lesson format needed to support it.
Making the lesson
fit the learner.
A parent wanted two uninterrupted hours of tutoring. I recommended shorter stretches with breaks. We both wanted the child to improve. We disagreed on how.
Start with the outcome the parent wanted.
The parent wanted their child to improve. More lesson time looked like a way to get there. I thought shorter stretches of focused work, with breaks, would work better for their child.
I explained the recommendation. Simply saying no would not address their concern. Simply saying yes would give them the format they asked for without doing what I thought the child needed.
Understand what mattered to each family.
A parent’s hopes and a child’s confidence were part of the conversation. Each family was different. I couldn’t use the same explanation or lesson format for everyone.
Children didn’t always have the words to explain what was wrong. I had to watch how they worked, notice where they lost focus or confidence, and change the lesson. Then I had to make that approach clear to the parent.
Stand behind the recommendation.
Some parents stopped the lessons. I accepted that rather than teach in a way I believed would work against their child. Some later returned after their children struggled elsewhere.
The difficult part was holding the recommendation without treating the parent as the problem. We wanted the same thing. I had to respect their concern while being honest about what I thought would work.
What I use in business
A buyer, a user and the people delivering a product can want different things from it. I work out where those needs overlap, where they conflict and what a decision would cost each person. In Toku, that means supporting the parent and tutor without losing sight of the child doing the practice.
Choosing what
to build next.
A leaderboard sounded good for Toku. That didn’t tell me whether it was the most useful thing to build next. I needed a reason beyond liking the idea.
What are we trying to change?
I start with that question. A leaderboard might encourage practice, but it isn’t automatically what a child needs most. I need to know the problem before I commit time to the feature.
Turn unknowns into testable questions.
What do we know? What would have to be true? What do we need to find out? If one answer depends on another, I break that down too. I keep going until there’s something I can test or do.
I aim to get that first map together within a day or two, depending on the problem. Then I build experiments around the assumptions that could change the decision.
Use the result to decide.
I want to know what would make me continue, change direction or stop. If every result would lead me to build the same thing, I haven’t designed a useful test.
Where the time goes
Every feature takes time from something else. I use the evidence to decide which work deserves it, then build and check whether it helped.
Serving the people
who need it.
I used to put too much weight on how many people could use a product. I changed my mind when I recognised what serving everyone could do to the product itself.
Start with people who feel the problem.
I used to be drawn to how broadly a product could apply. But trying to accommodate everyone can leave the people who need it most with something that no longer works well for them.
I now start with who feels the problem. What are they trying to do? What is getting in the way? What would make this worth using?
Choose features that strengthen the product.
Someone can have a real need without it becoming our next feature. I look at what serving that need would require and what would change for existing users.
Otherwise, every conversation adds another exception. The product gets harder to explain, harder to build and harder to use.
What changed
I still want to reach more people. I put less weight on the size of a possible market and more on how well the product serves the people who need it now.
Not everything
has to be work.
I make 3D art, experiment in the kitchen and get out for martial arts, running and rugby. I’m always up for doing more of it with other people.
Art, food experiments and sport.
3D art / mixed media / collaborationLet’s make
Let’s make
something.
I like 3D art and mixed media: combining forms, materials and techniques to see what happens.
I’d love to work with other artists and makers. We can start with an idea, an experiment or something neither of us knows how to make yet.
Opens your email appFrom scratch / fermentation / experimentsI like making it
I like making it
from scratch.
Fermentation is a particular favourite. I like changing ingredients and processes, waiting to see what happens, then trying again.
Plenty of attempts go wrong. That’s part of it. I’m up for swapping experiments or making something together.
Opens your email appMartial arts / running / rugbyI’m up for
I’m up for
getting out.
I enjoy martial arts, running and rugby. A training session, a run or a game is a good reason to meet.
Send me a note about what you’re doing. I’d be glad to join.
Opens your email appsee you
out there.
out there.