Daniel Oreofe and his co-founder are building the infrastructure layer that every AI application need to survive production.
They’re solving a simple to describe yet brutal to experience problem: the moment your AI product hits real users, everything you didn’t think about becomes a problem. Data leaks. Costs spiral. Zero visibility.
One bad response away from a disaster. Most teams discover this too late. Cencori sits between your application and your AI model, handling security, observability, and multi-provider routing automatically, from your first request. 160+ developers are already building on it.
In this interview with Habeeb Ajala, Daniel discusses their journey building Cencori, a complete infrastructure platform for building intelligent products.
Excerpts:
Can you briefly introduce yourself and what you are currently building?
I’m Daniel Oreofe, and I’m building two things simultaneously.
The first is Cencori, a complete infrastructure platform for building intelligent products. We’re building the layer that every AI product runs on: unified model routing, real-time security, persistent memory, compute, billing, deployment. The full stack. Incorporated in Delaware, built from Lagos. Cencori is live, developers are building on it today, and we’re only beginning.
The second is Dr. CV, under the Belycan brand. It’s a full end-to-end pipeline for job seekers. Upload your CV, get an honest recruiter-grade diagnosis in 30 seconds, then move through the rewrite, job matching, cover letter generation, interview prep, and application tracking, all the way to an offer.
Both companies sit on the same conviction: AI needs to be useful, not theoretical, and in Africa specifically, needs to be deployed honestly, affordably, and with a clear understanding of the environment it’s running in.
What first pushed you into technology or entrepreneurship?
I’ve always been drawn to systems, how they’re designed, where they break down, and why the people most disadvantaged by broken systems usually had the least hand in designing them.
What pushed me into entrepreneurship specifically was watching capable people around me get locked out of opportunities they deserved. CV formats built for Western hiring pipelines. Enterprise AI tools that assume infrastructure that doesn’t exist here. Job boards posting roles that were already filled. The gap between what technology promised and what it delivered in this context felt enormous and solvable, not inevitable. That gap is what I’m working on.
Was there a specific moment or problem that led you to start this journey?
For Cencori, it was pattern recognition. Every team building an AI product was solving the same seven infrastructure problems before writing a single line of product code: routing, rate limiting, cost control, billing, compute, memory, deployment. Each one a separate tool, separate contract, separate engineering sprint. Six to nine months of runway burned on infrastructure that has nothing to do with the actual idea.
Nobody had unified it. Not truly. There were gateways that didn’t bill. There were billing tools that didn’t route. There were compute providers with no security layer. Every serious builder was duct-taping the same fragmented stack together and calling it infrastructure.
That felt like the wrong answer to an enormous question. So we started building the right one.
For Dr. CV, the moment was simpler: I watched someone genuinely qualified get filtered out of a role because their CV wasn’t formatted the way an ATS expected. The skill was there. The document didn’t reflect it. That felt wrong and fixable.
What was the earliest version of your idea, and how has it changed since then?
Cencori started as a consulting practice advising companies on AI strategy. We moved away from that quickly when we realized that advice without implementation is an expensive opinion. Companies didn’t need someone to tell them AI was important. They needed someone to make it run. So we shifted to custom deployments, then recurring subscriptions, then managed infrastructure retainers. The evolution was from “we’ll help you think about this” to “we’ll build it, run it, and be accountable for it.”
Dr. CV started as a single diagnostic tool, upload your CV, get a score. But diagnosis without prescription is frustrating. So we built the rewrite layer, job matching across nine boards, cover letter generation, interview prep, and an application tracker. We realized that stopping at the CV meant we’d only solved step one of a ten-step problem. The product needed to care about whether you got the job.
What specific problem are you trying to solve, and who is most affected by it?
For Cencori: the problem is fragmented AI infrastructure. Right now, building an AI product means assembling eight different tools from eight different vendors before you ship anything. OpenRouter for routing. Pinecone for memory. Stripe for billing. Vercel for deployment. Each one a separate integration, a separate pricing model, a separate point of failure.
The teams most hurt by this are the ones who can least afford to waste the time: early-stage startups burning runway on infrastructure instead of product, solo developers who want to ship fast and can’t, enterprise teams who need compliance and security baked in and have to bolt it on themselves.
Cencori is one platform, one API key, every layer. You start with what you need today. The platform grows with you.
For Dr. CV: the problem is that talented people are losing at hiring before anyone has even looked at their skills. The job seeker in Lagos with five years of real experience getting filtered out by an ATS because their CV isn’t structured the way a recruiter in London expects it. The problem is the packaging, not the person, and we’re building the tool that fixes the packaging and then walks them through the rest of the process.
Why is this problem important in your local or African context?
Africa has a massive talent pool but a broken system for matching talent to opportunity, and a broken system for deploying the technology that could change how those businesses operate.
For AI infrastructure: companies that can’t deploy AI efficiently can’t scale efficiently. That compounds into slower hiring, slower growth, and a widening gap between businesses here and the ones competing against them globally. The cost gap matters too. When enterprise AI tools are priced for Western budgets, most companies here are priced out of the conversation.
For job seekers: almost everything about job hunting is calibrated for the wrong market. The CV advice online is built for American or European pipelines. The formatting norms, the language, the emphasis, none of it translates. And most people here can’t afford to pay a professional CV writer tens of thousands of naira for something that might not even work.
Both problems are solvable. They’re not being solved by the people building AI tools right now.
What makes solving this problem difficult in your environment?
Infrastructure, first. Unreliable power, expensive data, bandwidth constraints, all of it means AI solutions have to be leaner, more resilient, and more forgiving of connectivity issues than their Western counterparts. You can’t copy what works in California.
Trust, second. There’s real skepticism about AI tools in this market. Too many products have overpromised and underdelivered. Earning trust from enterprises and from job seekers who’ve been burned before takes time and requires delivering what you say you will, consistently.
Talent, third. Finding engineers who understand both advanced AI and African deployment contexts is genuinely hard. The ones who do are often being hired away by international remote roles. We compete for that talent every day.
Capital, last. Most funding infrastructure is still calibrated for US and European markets. The due diligence expectations, the ticket sizes, the relationship networks, they’re built around a different context. Building something real here before accessing that capital is possible, but it’s slower and more expensive on your own resources.
How did you take your first practical steps from idea to execution?
For Cencori, the first step was honest: I talked to people. Not “validated my idea” in the startup textbook sense. I genuinely asked enterprises what they’d already tried, what broke, and what they’d be willing to pay for something that worked. That converted into early advisory work, which gave us enough to cover the cost of the first deployment. Lagos has a surprising density of serious operators if you know where to find them.
For Dr. CV, we started with one working tool and no more. We didn’t try to build the entire pipeline first. We built the CV diagnostic, got it in front of real users, watched what confused them, watched what frustrated them, and kept asking the same question: what would make this move the needle for you? The answer kept coming back the same way. I don’t want to know my CV is bad, I want a better one. So we built the next piece. Then the next.
What has been your most significant milestone so far?
Shipping. The AI Gateway is live. Scan is live. Developers are building real products on Cencori’s infrastructure right now. For a platform company, the most significant milestone is always the same one: real users, real usage, in production. We’re there.
The second milestone is the Anthropic partnership. We’re an official Anthropic partner. That’s not a logo on a website. That’s a credibility signal that tells serious builders this infrastructure is real.
For Dr. CV: 1,000 users on the platform in barely two weeks of announcing. Real people who uploaded their real CVs and trusted our tool to tell them the truth. That number matters more to me than anything on a pitch deck.
What has been your hardest technical or operational challenge?
Building the security layer correctly. Jailbreak detection, PII masking, output scanning, all of it has to run in real time, before the request hits the model, without adding meaningful latency. You’re making a security decision in milliseconds on content you’ve never seen before. Getting that right required an architecture rethink, not better prompts.
The other hard problem is normalization. A hundred models across a dozen providers, each with different payload formats, different streaming behaviors, different error codes, and the developer should never feel any of that. One interface. Everything underneath is handled. That’s harder than it sounds.
For Dr. CV: the job matching layer. Aggregating across nine job boards, normalizing the data, and building relevance scoring that accounts for the African job market context, not Western keywords, has been technically intensive and is still being refined.
And honestly: building two companies simultaneously. Managing that context-switching without dropping either ball is its own daily challenge that no technical framework solves.
How are you currently funding or sustaining the work?
Bootstrapped. Both companies, entirely. Early revenue from Cencori deployments funds operations. We haven’t taken external capital yet, and that was a deliberate choice. We wanted to build something real before we walked into a room and asked for money. You can’t fake a working product.
We’re at the stage now where we’re actively building investor relationships and exploring funding programs. But the foundation was built on revenue. That matters to how we operate, and it matters to how we tell the story.
What is one decision you made that changed the direction of your journey?
Stopping consulting and starting to build. There’s a version of Cencori that never exists, one where we kept advising companies on AI strategy, charging comfortable retainers, and never shipping anything. That path was safer, but it was also a ceiling. The moment I said, “We’re not here to give advice, we’re here to build and own the infrastructure,” everything changed. The conversations changed. The revenue model changed. The ambition changed.
That one decision is why both companies exist today.
What failure or setback taught you the most?
An early client engagement where I overpromised on delivery timelines and underestimated the complexity of their existing infrastructure. We hit the deadline, barely, but at a cost to the team and to the quality of the initial deployment. We had to go back and fix things that should have been right the first time.
The lesson: honest scoping is a competitive advantage, not a weakness. Clients don’t want you to tell them what they want to hear on day one. They want you to still be around and accountable on day three hundred. I learned that slower than I should have.
If you could restart, what would you do differently from day one?
Start charging sooner. There’s a trap in early-stage building where you give things away: your time, your product, your advice, hoping it translates into traction or goodwill. Sometimes it does. More often it depletes your resources and trains the market to expect things for free. I learned that later than I should have.
I’d also have found my co-founders earlier. Building with people who fill your gaps is faster, more honest, and more sustainable than trying to be everything yourself.
Describe a typical working day for you right now.
The day starts before 8. Usually I’m already in my head, thinking about whatever is unresolved from the night before. Coffee, a few minutes where nobody can find me, and then I’m at the laptop. Mornings are for building: deep work, product decisions, writing, anything that needs full attention. Afternoons are for calls: clients, co-founders, partnerships. Lagos can be loud in the background of those calls, but people here understand that.
Somewhere in the middle of the day, if I haven’t moved, the dog reminds me. That walk is probably the most useful 15 minutes in the whole day.
Late evenings are for whatever didn’t fit elsewhere: reading, thinking through what tomorrow needs, processing the day without trying to fix everything. I don’t always succeed at that last part.
Where do you usually work from, and what does that space look like?
Home. A table, an ergonomic chair, and a laptop. No co-working theatrics, no café performativity. I’ve found I build better when I’m not performing productivity for anyone watching. The space is simple because complexity in the environment competes with clarity in the work.
What sounds, routines, or distractions are part of your daily building process?
Lofi music, almost always. It’s the right kind of background: present enough to crowd out distraction, not so present that it demands attention.
The dog is the best interruption I have. When I’ve been staring at the same problem for two hours and my brain starts to loop, he physically forces me outside. Fifteen minutes of walking and breathing different air, and most things resolve on their own.
My phone is the bad distraction. I haven’t fully solved that one.
Who are the people around you during a normal working week?
My co-founders. We’re building different things with shared conviction, and that relationship is different from a typical team. More honest, more argumentative, more generative.
My family, present even when they’re not in the room. I grew up watching capable people work hard and not get what they deserved from it. Building something that could change that equation is not abstract for me. It’s daily.
And then there are people around me who’ve never started anything, who watch what I’m doing and ask questions. I think I inspire them. I try to be honest about how hard this is, because inspiration that omits the difficulty isn’t useful.
What does a “stressful day” look like for you in practical terms?
Multiple meetings stacked on a calendar that was already full, preceded by a late night where I was still working on something at 1am. Starting the morning already behind before you’ve done anything.
Technical integration challenges hit differently though. When something doesn’t work and you can’t find why, and real users or real clients are on the other side of that brokenness, that stress doesn’t close with the laptop. It follows you.
What does a “good day” feel like when things are working well?
A testimonial. Somebody I’ve never met, using something I built, telling me it changed something for them. That’s the whole thing.
Also food. A good day has food in it. In early-stage startup life, that is not always guaranteed, and I say that only partially as a joke.
What keeps you going on difficult days?
My family. And survival. I’ll be honest about that. There is no safety net here. If I stop, I don’t have the luxury of landing somewhere soft. There’s no one coming to save me.
I know that’s not the romantic answer. But it’s the real one. The people building seriously in Nigeria often don’t have fallback options, and that creates a particular kind of focus I’ve stopped being embarrassed about.
What personal sacrifice has been necessary to keep this venture alive?
My savings. I spent them. There’s nothing left in the sense most people my age would expect to have by now.
I made the calculation that the opportunity was worth it, and I still believe that. But I want to be honest: it wasn’t a painless bet. It cost something real.
How has this journey changed how you see yourself?
I’ve built myself more than I’ve built the business. That’s the honest answer.
The most visible change: rejection doesn’t land the same way anymore. When a potential lead says no to a solution you’ve been refining for months, that’s a specific kind of hit. By the time you’ve absorbed enough of those, a girl saying no genuinely feels manageable. The thresholds shift.
What I’ve also found is a kind of self-trust that wasn’t there before. Not arrogance. There’s too much I still don’t know for that. But a real confidence that when something difficult is in front of me, I’ll figure it out or build through it. That wasn’t available before I started.
What kind of impact do you hope your work will have in the next 3 to 5 years?
Dr. CV: 500,000 job seekers in Nigeria, meaningfully helped. We hit 1,000 today. The gap is clear and I find that motivating rather than discouraging. The work isn’t done yet.
Cencori is the infrastructure platform that the intelligence era runs on. Not for one country, not for one continent. For every developer, every team, every company building intelligent products anywhere on earth.
In three to five years, the name “Cencori” means what “AWS” means when someone asks where their infrastructure runs. It means the serious builders chose it. It means the question of which AI infrastructure platform to use has a default answer.
The intelligence era needs one infrastructure company the way the internet era needed AWS and the mobile era needed Stripe. We intend to be that company. That’s not a vision statement. That’s the plan.
What is the next big step or ambition for your venture?
Capturing the African market: intentionally, methodically, with the depth of deployment that makes us hard to displace. Not a land grab. A foundation. If we build the infrastructure correctly, everything else compounds from it.
If someone is reading your story today, what do you want them to learn from it?
Build scared. Build unknown. Walai, the journey will unveil itself while you’re building.
Clarity is not a prerequisite for starting. Starting is how you get clarity. The people waiting to feel ready are waiting for something that doesn’t arrive before the attempt.
Is there anything about your journey that people usually don’t ask, but you think is important for them to know?
My mental health.
I’m not fine. I’ll say that plainly. There are a thousand things to do, a hundred of them urgent, and the list never gets shorter. It changes shape. I’m always busy. Always. The ambition is real, and the load is real, and I don’t think enough people building at this level are honest about the cost of carrying both.
I’m not asking for sympathy. I’m functional, I’m building, I’m still here. But if someone reads this and thinks “he has it all figured out,” I want them to know: I’m navigating it in real time too. That’s true for most founders I know who are being honest about their situation.
The appearance of certainty is often the hardest thing we build.
