Ten Skills I Gained Building Tech Companies (That I Wish I'd Learned Sooner)

I started my career as a developer. Coding was the easy part. The hard part, the part I had no training for, was everything that turns code into a company. The CIAM platform I built scaled to a billion users, then to a real revenue line, then to a real organisation. Every step taught me something I wish someone had handed me on a printout in year one.

These are the ten skills I actually use now, building my next company in AI and security. None of them came from a textbook. Most of them came from getting them wrong first.

1. Selling before you have anything to sell

Engineers underrate sales because they think it is about persuading people to buy things they do not need. It is not. Sales at the founder stage is about getting a stranger to spend 20 minutes of their day with your problem instead of theirs. That skill compounds across every part of the job: recruiting, fundraising, partnerships, customer development, design feedback.

The thing nobody tells engineers: you can practise sales without a product. Cold email a hundred CIOs about a problem you find interesting. Ask for 15 minutes. Take notes. By the time you have a product, you will know how to talk about it because you have already had a hundred conversations about the problem.

2. Listening as a hiring skill

Active listening is sold as a relationship-building trick. It is, but the higher-leverage use for founders is recruiting and customer discovery. Most founders interview by talking. The good ones interview by setting up a question and shutting up for as long as it takes.

I learned this the expensive way. The senior hires I regretted were the ones I out-talked in the interview. The ones who shaped the company were the ones who got 80% of the talking time, where I asked one question and listened to the next 45 minutes. People reveal themselves when they have space to.

3. Storytelling, not presenting

Slide decks are the form. Story is the substance. The investors who said yes to me did not say yes to a chart; they said yes to a narrative they could repeat to their partners on Monday. Same for the first 10 customers, same for the first 10 hires.

The story has four moving parts: (1) a clear before, (2) a sharp inciting incident that made you start, (3) a concrete after, and (4) a credible reason you are the one to build the bridge between them. Most founder decks have three of these and miss one. Usually number 4.

4. Conviction over confidence

Confidence is performative. Conviction is structural. The first is what you project; the second is what you actually believe based on the evidence you have gathered. Investors, customers, and senior hires can tell the difference within about ten minutes.

The way to build real conviction is unglamorous: spend more time in the problem than anyone else. Not more time writing decks. More time on calls with users, more time with the data, more time with the failure modes. If you have done that, the confidence shows up on its own. If you have not, no amount of LinkedIn posting fills the gap.

5. Writing as the highest-leverage skill

The single skill I would tell my younger self to invest in earlier is writing. Not blog-post writing. The kind of writing that goes into one good update, one clear PRD, one well-aimed cold email. Clear writing is clear thinking made visible, and it scales: the same paragraph reaches one person or a million.

The practical version: every Friday, write a 200-word update to your team or your investors. Make it specific. Numbers, decisions, what changed, what you got wrong. After 50 weeks, you will write 10x better than you did at week one, and you will have a memory of the year that is more honest than anything else you produce.

6. Distribution beats product quality (most weeks)

Engineers want this not to be true. It still is. The difference between a great product nobody hears about and a good product everyone hears about is the second one wins the market every time. The product can catch up; the audience cannot.

The implication for tech founders: spend 30-50% of your time on the channels that move customers to you, not on the product itself. SEO, LinkedIn, podcasts, communities, founder-led sales, partnerships. None of these are glamorous. All of them compound. How companies achieve AEO and GEO and the GEO market research cover the 2026 version of this for AI-driven search.

7. Saying no as strategy

Nothing kills a young company faster than saying yes to too many things. Every feature request, every partnership conversation, every adjacent market that looks interesting. Each of those is a small yes that adds up to a fragmented, mediocre version of the company you set out to build.

The framing that helped me: "What am I giving up by saying yes to this?" If the answer is fuzzy, the answer to the new opportunity is no. Founders who learn this early build sharper companies. Founders who learn it late build big, brittle ones.

8. Hiring is the multiplier you cannot opt out of

For the first 30 employees, every hire either doubles your velocity or halves it. There is no neutral hire. I have made expensive versions of both. The signal that matters is not the resume; it is whether they raise the bar for the people already there.

Two specific habits that helped me later: (1) reference checks on every senior hire, conducted personally, not delegated. Three calls, 30 minutes each. (2) Hire the person whose questions in the interview are sharper than your answers, not the person who has the right answers to all your questions.

9. Capital as a tool, not a trophy

Raising money is not the achievement. It is the obligation. Each round resets the clock and raises the bar for what counts as success. The founders who do this well think about capital as fuel for a specific objective, not as validation.

The most useful question I have learned to ask: "What do I need to prove in the next 18 months to make the next round inevitable?" Then I work backwards from that to the capital I actually need. Raising too little is dangerous; raising too much is more dangerous, because it teaches the team to spend at the wrong rate.

10. Staying calm when it is loudest

Every founder gets the call that the customer is leaving, the engineer is quitting, the bank account is short, the investor is wobbling, all in the same week. The job in that moment is to be the person whose voice goes down, not up. Whose pace slows, not speeds. Whose decisions get more precise, not more reactive.

This is not a personality trait; it is a practised response. Sleep, exercise, a handful of senior advisors you trust, and a stubborn refusal to make irreversible decisions when emotional are the four things that built it for me. The companies that survive the bad weeks have founders who could be in the room for them.

The meta-skill underneath all ten

The thread running through every one of these is the willingness to be specifically bad at something publicly while you get better. Founders who stay anonymous, or who refuse to do anything they are not already good at, plateau quickly. Founders who write the first bad blog post, give the first bad pitch, do the first bad sales call, and then do it again the next week are the ones who compound.

None of these are the skills a computer-science degree teaches. None of them show up on a job description. All of them are why some founders go from "good engineer with an idea" to "company that exists at scale" and most do not.

Adjacent reading on guptadeepak.com

For the longer-form founder narrative: my journey. For the practical playbook: the complete guide to setting up your US tech startup and the San Francisco tech scene guide. For the strategic frameworks I use across AI and B2B SaaS: the AI revolution toolkit.

FAQ

Which of these ten matters most in the first year?

Selling and writing. Selling because nothing else matters if no one wants what you are building. Writing because it forces you to think clearly enough about what you are building to explain it to other people.

How do you teach yourself sales as an engineer?

Cold email a hundred people with a question, not a pitch. Take their answers seriously. By the time you have a product, you will know what language your buyers actually use, which is the entire job.

Is there a skill I should explicitly hire for instead of learning myself?

Yes, the one that is most opposite to your natural mode. Engineers should hire a strong commercial co-founder or first sales hire early. Commercial founders should hire a deeply technical co-founder. The temptation is to hire your own image; resist it.

How long does it take to build these skills?

Each one takes between 6 and 24 months of deliberate effort. The good news is they compound: the writing makes you a better seller, the selling makes you a better listener, the listening makes you a better hirer. You are not building ten skills in series; you are building one network of skills in parallel.

What's the most overrated founder skill?

Being a visionary. The founders I admire most are unglamorously specific about what they are building, why, and for whom. The visionary frame is mostly used by founders who have not yet got specific.

Where do most technical founders fall short?

Distribution. They build something good, then assume the world will find it. The world will not. Distribution is the part of the job they do not enjoy, which is precisely why investing in it is high-leverage; most of their competitors are also avoiding it.