Results
- 10+go-lives of XNETT at PremiumTrust BankNigeria's Fastest Growing Bank. 6 branches then, 37 now
- $300K+in 3 monthsA manual-first launch I proposed at Revolve
- 55% to under 15%failing testsTwo client apps, MySirigu and Tra-Vu, at LarvaBox
- 2,000+users on YadsaleA marketplace I help run, grown as a team of 9
In short
A bank, a payments startup, two client apps and a P2P marketplace. I'm the False 9: between the people with the problem and the system meant to solve it.
“If you do everything, you'll win.”
Every role handed me something. I kept all of it.
Experience
Technical Product and Operations Manager, YadsalePeer-to-peer marketplace with escrow payments (SafePay)
Feb 2025 to present. Part-time. Co-lead a cross-functional team of 9.
- Removed the account wall: turned funnel and session-recording insights into guest checkout and guest offers, so buyers no longer need an account to buy.
- Found a sharp drop-off at OTP verification during sign-up, traced it to failed SMS delivery on certain numbers, and the team shipped a WhatsApp OTP fallback.
- Set up PostHog, GA4 and Microsoft Clarity, and built funnels from first visit to purchase.
- Turn business and user problems into build-ready specs with acceptance criteria: revenue reporting, coupon spend tracking, role-based admin permissions.
- Investigate issues across checkout, offers, payments, escrow and refunds using Postman, SQL and dev tools.
- Segmented users with SQL for campaigns reaching 1,500+ people, and led campaign direction for the interstate delivery launch.
Platform context, not personal credit: 2,000+ users. The escrow flow was designed as a team. Stack: Next.js, Node.js, PostgreSQL, Paystack.
MSc thesis, built solo, MSc Big Data Management and Analytics, Griffith College Dublin
Jan 2025 to Jun 2026
- A Google Cloud pipeline that turns informal buying requests on X ("who sells?") into demand intelligence.
- X API v2 to Cloud Storage, Python (rule-based NLP, quality scoring), BigQuery and Looker Studio, orchestrated by Airflow on Cloud Composer.
- 1,845 actionable demand signals across 24 categories, from 3,032 validated records.
Technical Product Manager and QA Engineer, Revolve Inc.Cross-border payments startup on stablecoin rails
Oct 2024 to Sep 2025. Contract, remote. Corridors: USD to NGN, and USDT to NGN or USD.
- Proposed a manual-first go-to-market that processed $300K+ in transactions within three months, while the platform was still delayed: customers verified through Prembly, then served directly over WhatsApp.
- Brought in Prembly for customer KYC, and joined the evaluation calls with KYC providers.
- Led product planning for the platform's first version and ran technical project management.
- Sole QA: tested pre-launch builds and APIs.
QA Engineer and Technical Project Manager, LarvaBoxProduct consulting firm for early to mid-stage companies
Jun 2024 to Jun 2025. Contract, remote. Team of 5 to 6.
- Cut the test failure rate from 55% to under 15% by flagging recurring defect patterns until they stopped recurring.
- Sole QA and technical project manager for two client apps: MySirigu, for gifting and crowdfunding, and Tra-Vu, for shared commuting.
- Reviewed every PRD with the product manager (the project lead) and made build decisions together on both apps. Set up the Tra-Vu team with them and contributed to its architecture decisions from the early stage.
- Owned day-to-day delivery with the engineers: kept tasks current, tested every push, and held each release to good UX before deployment.
- Wrote 300+ test cases while MySirigu was built, tested 30+ of its builds through launch and 17+ Tra-Vu releases on TestFlight and Google Play, and ran client demos and UAT for both.
- Validated Paystack and Stripe integrations for users in Nigeria, the UK and the US.
- Verified REST APIs in Postman. Diagnosed websocket and Firebase authentication defects.
- After duplicate records reached production, partnered on a clean-up script and added idempotency checks to regression testing. No duplicate-entry incidents afterwards.
- Wrote product documentation and the architecture flowchart used in the client handover.
Platform context, not personal credit: MySirigu onboarded 150+ users in its first month; Tra-Vu's MVP shipped on time with every feature in scope, for Nigeria, the UK and the US.
Organising committee, speaker management, TEDxAbuja, volunteer
Jan 2023 to Jan 2024. Theme: "Bottom Up".
- With the speaker team, prepared 8 speakers: rehearsals, review against TED content guidelines, and travel coordinated around the schedule.
- Co-built pre-event, event-day and post-event trackers with the 29-person committee.
- Contributed to partnership decks and sent invitations to prospective dignitaries.
- Managed on-site registration for walk-in guests at an event of 118+ attendees.
- Co-developed the theme narrative anchored on Ad astra per aspera ("through hardships to the stars").
Technical Consultant, Deployment and Integration, FASYL Technology Group, at PremiumTrust BankPan-African banking and payments technology group (40+ institutions, 13 countries), embedded at a bank
Nov 2022 to Feb 2024. Hybrid, Lagos. 15 months embedded at the bank. Earlier title: SQL Developer.
- Took XNETT, FASYL's clearing and transaction platform, live 10+ times across PremiumTrust Bank's 6 branches: electronic funds transfer and cheque truncation, connected to the national clearing house and posting into Oracle FLEXCUBE.
- Was the bank's main point of contact for the final 9 months, and demoed every go-live and change request to its stakeholders and project managers.
- Turned feedback from cashiers, service managers and the clearing team into product changes: compliance-controlled transaction limits, direct debit as a new clearing instrument, and a clearer clearing interface.
- Validated transactions end to end into FLEXCUBE, including scanned cheque image data.
- Handled L2 support, resolving production incidents in 10 to 60 minutes depending on severity.
- Wrote deployment, migration and rollback documentation that let successor consultants deploy on their own.
- Customised the E-Alert messages PremiumTrust customers receive after every transaction.
PremiumTrust opened in 2022 and was named Nigeria's Fastest Growing Bank 2022 (BusinessDay BAFI Awards). Its transfers, cheques and payments ran on XNETT and XNETT Pay, and XNETT kept clearing as the bank grew from 6 branches to 37.
Technology Analyst (internship), Ensalvia Technology ConsultingTechnology consultancy
Jul 2022 to Oct 2022
- Supported current-state assessments for enterprise clients, finding operational gaps and recommending system-level improvements.
- Contributed research to a digital transformation proposal for a state government.
- Ran the options analysis for a fleet management system and a competitor analysis for a peer-to-peer product.
- Gathered requirements from stakeholders across client organisations.
Case studies
Bringing XNETT live at PremiumTrust Bank
- The story
- A bank that can't clear can't work. Cheques sit. Transfers stall. And PremiumTrust, new in 2022, was named Nigeria's Fastest Growing Bank that year. FASYL builds the systems behind banking and payments for 40+ institutions across 13 African countries. PremiumTrust's transfers, cheques and payments ran on its platforms, XNETT and XNETT Pay: clearing through the national clearing house and into Oracle FLEXCUBE, the bank's core system. Building it was FASYL's job. Making it work inside the bank, for the people who'd use it every day, was mine. A new bank opening branches at that pace can't stop to wait for its systems. So for 15 months, that's where I sat. For the last 9, I was the bank's main point of contact.
- How it works
- XNETT Pay moves transfers and collections in real time, with maker-checker approvals before anything leaves. A transfer starts in EFT. A cheque is scanned at the branch in CTS, and its image is stored with its data. XNETT exchanges it with the national clearing house. The cleared item posts into FLEXCUBE, and it has to match exactly: cheque number, customer name, amount. Then the customer gets an E-Alert.
- What I did
- Took XNETT live 10+ times across the bank's 6 branches, and demoed every release to the bank's stakeholders and project managers before it went out. Listened to the cashiers, service managers and clearing team, and turned what they said into product: transaction limits compliance could control themselves, direct debit as a new clearing instrument for the bank, and a clearer interface for the clearing team. Built those changes in Flask (Python), Angular, TypeScript and Oracle SQL. Checked every transaction from XNETT into FLEXCUBE, down to the scanned cheque images in the database. When direct debits failed in production, traced it to malformed XML from empty database results, fixed it, and verified it before the next settlement window. Ran L2 support, resolving incidents in 10 to 60 minutes by severity. Once, that meant getting the head branch's cheque scanner back up. Customised the E-Alert messages PremiumTrust customers get after every transaction. Wrote the deployment, migration and rollback docs, so the consultants after me could set up servers and deploy without me.
- What changed
- 10+ go-lives, as the bank grew from 6 branches to 37. A bank growing that fast needs clearing and payments that keep up. XNETT kept up. The platform was FASYL's. Making it work at the counter, and keeping it working, was mine.
Two apps, one QA
- The story
- Two client apps. Three countries. One QA. Me. MySirigu lets people raise money for birthdays, weddings, causes and campaigns, in Nigeria and the US. Tra-Vu matches drivers who have empty seats with commuters on the same route, in Nigeria, the UK and the US. LarvaBox, a product consulting firm for ambitious early to mid-stage companies, built both for its clients, and I tested every build of both. On both, I was the technical project manager, working beside the product manager, who led. More than half our test runs were failing, and it was the same bugs, again and again. Then duplicate records reached production.
- How it works
- Every build goes through regression testing, then client demos and UAT. Payments run through Paystack and Stripe, for users in Nigeria, the UK and the US. Every API gets checked in Postman before anything ships, then builds go out through TestFlight and Google Play.
- What I did
- Reviewed every PRD with the product manager, on both apps, and made the build decisions together. Set up the Tra-Vu team with the product manager, and was part of its architecture decisions from the early stage. Kept the engineers' tasks current, tested every push, and held each release to good UX before it went out. Wrote 300+ test cases while we built MySirigu, then tested 30+ of its builds through launch, and 17+ Tra-Vu releases on TestFlight and Google Play. Demoed to both clients with the product manager, and worked with them through UAT. Flagged the same patterns of bugs back to the mobile developer, every time, until they stopped coming back. Tracked down websocket and Firebase authentication bugs with the backend and mobile engineers. When the duplicates hit, helped write the script that cleaned them up, then added idempotency checks to every regression run. Wrote the product docs for both apps, and drew the architecture flowchart the client's own developers took over with.
- What changed
- 55% to under 15% failing tests. No duplicate records since. The fixes were the developers'. Flagging it until it stuck was mine. And both apps shipped: MySirigu onboarded 150+ users in its first month, and Tra-Vu's MVP landed on time with every feature in scope.
The customers were ready. The platform wasn't.
- The story
- Revolve moves money across borders on stablecoin rails, between USD, USDT and NGN. Business customers wanted in. The product wasn't ready for them. So I proposed we stop waiting. Verify customers through Prembly, our KYC partner, then serve them by hand, directly on WhatsApp. The founding team said yes.
- How it works
- A customer is verified through Prembly. We serve them directly on WhatsApp. Money moves on stablecoin rails: USD to NGN, and USDT to NGN or USD.
- What I did
- Proposed the manual-first launch. Brought in Prembly for KYC, and sat in on the calls choosing a provider. Led planning for the platform's first version, and ran its technical project management. Tested every pre-launch build and API, as the only QA.
- What changed
- $300K+ moved in three months. That's volume, not revenue. The idea was mine. The operations team made it work.
Every extra step is a door
- The story
- Every extra step between a buyer and a purchase is a door someone walks out of. Yadsale is a peer-to-peer marketplace in Nigeria: your local market, but safer. You can haggle, and SafePay holds the buyer's money until the item arrives. I co-lead its 9-person team, across tech and marketing. Buyers used to have to create a profile and verify by SMS before they could buy anything. Some never made it through. The data told us where. The recordings told us why.
- How it works
- PostHog, GA4 and Microsoft Clarity follow a buyer from first visit to purchase. Funnels show where people leave. Session recordings show why. Payments, escrow and refunds run on Paystack, through SafePay.
- What I did
- Set up the analytics, and built the funnels. Found a sharp drop at SMS verification, and traced it to texts failing on certain numbers. The team shipped a WhatsApp fallback. Turned what we learned into guest checkout and guest offers. Write the specs, from revenue reporting to coupon tracking to admin permissions, and catch edge cases before release, like a buyer ordering more than a seller has. Dig into payment, escrow and refund issues with Postman and SQL. Split our users into sellers, buyers, both and inactive with SQL, for campaigns reaching 1,500+ people, and led the campaign for our interstate delivery launch.
- What changed
- No account needed to buy. The platform, not me alone: 2,000+ users.
Purchase intent signals on X (MSc thesis)
- The story
- Every day, someone on X asks: who sells? That's demand. Real, specific, and scattered across thousands of posts no seller will ever read. So for my MSc, I built something to read them.
- How it works
- X API v2 pulls the posts into Cloud Storage. Python sorts and scores them with rule-based NLP. The results land in BigQuery and Looker Studio. Airflow on Cloud Composer runs the whole thing.
- What I did
- Built the pipeline, end to end, on my own.
- What changed
- 1,845 buying signals across 24 categories, from 3,032 validated posts. 67.2% were people simply trying to find something.
Eight speakers, one stage
- The story
- Eight speakers. One stage. One day. The theme was "Bottom Up", anchored on ad astra per aspera: through hardships to the stars. With the speaker team, my job was to get eight people ready, within TED's guidelines, and in the right city on the right day.
- How it works
- Rehearsals and reviews against TED's content guidelines. Travel planned around the schedule. Trackers for before, during and after the event, shared across the committee.
- What I did
- Prepared the speakers, with the speaker team. Built the trackers, with the rest of the 29-person committee. Ran registration for walk-in guests on the day.
- What changed
- 8 speakers on stage, for 118+ people. A team of 29 made that happen. I was one of them.
Education
- MSc Big Data Management and Analytics, Griffith College Dublin, Jan 2025 to Jun 2026. Thesis: a Google Cloud pipeline that turns informal buying requests on X into demand intelligence.
- BSc Computer Science, Babcock University, 2018 to 2022.
- Read the thesis article
Skills
Payment Systems, REST APIs, Solution Implementation, Product Demonstration, SQL, Python, PostgreSQL, Postman, Webhooks, OAuth, Paystack, Stripe, Clearing and Settlement, KYC, ISO 20022, Idempotency, PostHog, GA4, Tableau, Google Cloud (BigQuery, Airflow), Linux.
Certification: ISTQB Certified Tester Foundation Level.