What should you ask a potential technical co-founder?
Short answer
Ask a potential technical co-founder questions that make them show how they decide, not what they know: when their last system failed and why, what they would refuse to build, how much time they can commit and when, what they need to earn, and how they handle being overruled. Good answers are specific, admit trade-offs and mistakes, and connect technology to customers. Rehearsed certainty and vague commitment are the warning signs.
By Dr. Dulan Dias, co-founder and CTO of Tellme AI
Key facts
- Chapter 16 of From Model to Product suggests that instead of asking whether a model works, it may be more useful to ask when it fails.
- In the Stack Overflow Developer Survey 2026, 73% of respondents said they use coding assistants and agents daily.
- Under section 11 of the Copyright, Designs and Patents Act 1988, the author of a work is its first owner unless it is made by an employee in the course of employment.
- Section 3 of the Copyright, Designs and Patents Act 1988 defines a literary work to include a computer program, so source code is protected by copyright.
- Section 175 of the Companies Act 2006 gives every director a duty to avoid conflicts of interest.
- The ICO says that if a personal data breach is likely to pose a risk, the organisation must notify it as soon as possible and, where feasible, within 72 hours.
- The NCSC describes Cyber Essentials as the minimum standard of cyber security the Government recommends for organisations of all sizes.
- At EchonLabs, Dr. Dulan Dias built a Janashakthi Life digital insurance platform designed to stay usable over 2G and 3G, which was adopted by 15+ financial institutions across 300+ branches.
Definitions
- Technical judgement
- The ability to choose the simplest approach that meets real constraints such as cost, time and risk, and to know what not to build.
- IP assignment
- A written agreement that transfers ownership of intellectual property, such as code, designs or inventions, from the person who created it to the company.
- Shareholders' agreement
- The contract between a company's shareholders, usually alongside its articles of association, covering vesting, leaver terms, decision-making and what happens when someone leaves.
Technical judgement: how do they decide what to build?
Ask about decisions and failures, not technologies. Chapter 16 of From Model to Product, Working in Teams, suggests asking when a model fails rather than whether it works; the same question works on people.
| Question | A good answer sounds like |
|---|---|
| Tell me about a system you built that failed in production. What happened? | A specific incident, what they missed, what they changed afterwards and what it cost |
| What constraint decided the design of something you built? | A concrete limit such as bandwidth, budget, latency or regulation, and the trade-offs it forced |
| What would you leave out of our first version? | A short list, with reasons tied to users and risk rather than fashion |
| How would you build our MVP in a quarter of the time? | Cuts that keep security and data safe, and a plan to revisit what was cut |
| How do you use AI coding tools, and how do you check their output? | Regular use with a clear review habit: tests, reading the changes, and knowing where they do not trust it |
What does a strong answer to a judgement question sound like?
A strong answer names the constraint, the choice it forced and what the choice cost. From practice: at EchonLabs, Dr. Dulan Dias built a Janashakthi Life digital insurance platform designed to stay usable over 2G and 3G, and it was adopted by 15+ financial institutions across 300+ branches. The connection speed, not the framework, decided most of that design. Listen for answers with that shape.
AI tools are now part of the answer. In the Stack Overflow Developer Survey 2026, 73% of respondents said they use coding assistants and agents daily, so the useful question is not whether a candidate uses them but whether they can tell when the output is wrong.
Ownership: will they own outcomes or tasks?
These questions show whether a candidate treats the product's success as their job or only the tickets assigned to them.
| Question | A good answer sounds like |
|---|---|
| Who was responsible the last time your product went down at night? | They were, or they know exactly who was and why; they describe the fix and the follow-up |
| What did you do the last time a priority you disagreed with was chosen? | Committed to it, said what they would watch, and raised it again with evidence |
| What part of a company outside engineering have you had to learn? | Sales, support, finance or compliance, with an example of how it changed a technical choice |
| How would you know, six months in, that you were doing this job well? | Outcomes such as users, reliability and delivery, not hours or lines of code |
Values: what will they do with users, data and the truth?
Values show up in how someone treats users' data and bad news, so ask about concrete situations rather than principles.
| Question | A good answer sounds like |
|---|---|
| What would you refuse to build, even for a paying customer? | A real line, with the reason, not a slogan |
| How would you handle a data breach in our first year? | Contain it, assess the risk, tell the right people honestly, fix the cause, and know that a breach likely to pose a risk must be reported to the ICO, where feasible within 72 hours |
| What security would you put in place before our first customer? | A short practical list such as access control, secrets, backups and updates; many will name Cyber Essentials |
| How would you explain to a customer what our AI cannot do? | Plain language, real failure cases and what the product does when it is unsure |
| When did you last tell a manager or customer something they did not want to hear? | A specific example, told without blame |
Commitment: time, money and the rest of their life
Commitment questions test whether the candidate's life allows the role, not only whether they want it.
| Question | A good answer sounds like |
|---|---|
| When could you go full time, and what has to happen first? | A date or a clear trigger, such as a funding round, that they have thought through |
| How many months could you go without a salary? | A number, with the household reasons behind it |
| What other jobs, projects or boards are you involved in? | A full list and what they would drop; if you are both directors, each of you has a legal duty under section 175 of the Companies Act 2006 to avoid conflicts of interest |
| Where will you live and work for the next three years? | A clear answer, including any visa or family constraints |
| What would make you leave? | An honest answer such as health, money or a breakdown of trust, not 'nothing' |
Money: equity, pay and who owns the work
Money is easier to discuss before anyone has an emotional stake in a number; technical co-founder equity covers how splits and vesting work in the UK.
The IP question matters in law as well as in principle. Under the Copyright, Designs and Patents Act 1988, a computer program is a literary work and its author owns it unless it is made by an employee in the course of employment, so code a co-founder wrote before joining stays theirs until they assign it.
| Question | A good answer sounds like |
|---|---|
| What equity split would feel fair, and why? | A reasoned view tied to commitment and timing, open to discussion |
| Will you accept vesting and leaver terms on your shares? | Yes, on the same terms as the other founders |
| Will you assign everything you build, including what you have built so far, to the company? | Yes, without hesitation |
| What salary do you need, and from when? | A figure and a timetable, not 'whatever works' |
Conflict: how do they disagree?
Every founding team disagrees; what matters is whether a candidate can disagree, decide and move on without lasting damage.
| Question | A good answer sounds like |
|---|---|
| Tell me about a serious disagreement with a co-founder or manager. How did it end? | Specific, fair to the other person and clear about what they learnt |
| If we disagree on what to build, who decides? | A proposed rule, such as the CEO deciding priorities and the CTO deciding architecture, with a way to escalate |
| What should happen if one of us wants to leave? | Agreed leaver terms, a handover and no surprises for customers or staff |
| What would you want me to do if I thought you were underperforming? | Tell them early and directly, with examples |
How should you run the conversations?
Run them as a series of working conversations, not a single interview:
- Spread the questions over several meetings in different settings, including at least one long working session.
- Answer every question yourself as well; they should be judging you too.
- Write down the answers you agree on. Equity, vesting, IP, decision rights and leaver terms belong in a shareholders' agreement drafted by a solicitor.
- Check what you can: talk to people they have worked with, look at systems they built, and look up past directorships on Companies House.
- Then work together on a short, real project before committing; how to find a technical co-founder in the UK explains how to run one.
Frequently asked questions
How many conversations should you have before choosing a co-founder?
Several, over a few weeks at least, and ideally a real piece of work together before agreeing anything. One long dinner tells you whether you get on; it does not tell you how someone behaves when a deadline slips or a customer complains. Cover every group of questions on this page, and do not skip money and conflict because the conversation is going well.
Should you ask a potential co-founder for references?
Yes, and offer yours in return. Ask to speak to two or three people who have worked closely with them, ideally including someone they managed or disagreed with. Ask what the person was like under pressure and whether the reference would work with them again. A candidate who cannot name anyone who has seen their work up close is a risk, however well the interview went.
What should a technical co-founder ask a non-technical founder?
Mostly the same questions in reverse, plus a few of their own: what evidence shows customers want this, how the founder will sell before the product is ready, how much money they can put in or raise, and how they make decisions about things they do not understand. A technical co-founder should also ask who will own the relationship with customers, because that decides what gets built.
Should you give a potential technical co-founder a coding test?
Not a standard test. A strong engineer will reasonably decline an interview exercise from someone asking them to be a partner. A short real project together is better, because it shows how they work, explain decisions and take feedback. If you cannot judge code yourself, ask an experienced engineer you trust to review what they build or to join one technical conversation.
How do you check a technical co-founder's past work?
Ask them to walk you through something they built, using the real system, code or architecture where they are allowed to share it. Listen for what they would change, not only what went well. Look at public work such as open-source contributions or talks, speak to people who worked with them, and check any past UK directorships on Companies House, which shows current and resigned officers free of charge.
Should co-founders discuss what happens if the company fails?
Yes, while failure still seems unlikely. Agree how long you will keep going without income, what signals would make you stop or change direction, what happens to the code and any customer data, and how you would treat staff and investors. Founders who have discussed failure calmly in advance decide better when it comes close, because the hardest conversation has already happened.
Do you need a lawyer before agreeing terms with a co-founder?
You need one before signing, not before talking. Agree the principles between yourselves using questions like these, then have a solicitor turn them into a shareholders' agreement and articles covering vesting, leaver terms, IP assignment and decision-making. The cost is small next to a dispute over shares, and investors will read these documents during due diligence.
Sources
- Stack Overflow Developer Survey 2026, AI
- Copyright, Designs and Patents Act 1988, section 11
- Copyright, Designs and Patents Act 1988, section 3
- Companies Act 2006, Part 10, Chapter 2: general duties of directors
- ICO, Personal data breaches
- NCSC, Cyber Essentials overview
- GOV.UK, Get information about a company
About the author
Dr. Dulan Dias is co-founder and Chief Technology Officer of Tellme AI in Belfast; founder of EchonLabs, grown from 2017 to more than 60 people before its acquisition in 2026; holder of a PhD in AI and data science from Queen’s University Belfast; and the author of From Model to Product.