Just went live with our facial recog app! • Python AI powered the core model • React frontend streams video & UI • NestJS backend
My jam? Building the server & syncing our AI & React teams from dev → prod. Let’s go!
@ToksNet
Spent the last few weeks building Zephlo an inventory/procurement system with a twist: you can define entirely new business objects (Customers, Patients, Vehicles, anything) from the UI, with custom fields and workflows. No code.
Here's the demo 🧵
The kicker: switch the whole system from "restaurant" to "clinic" different departments, different fields, different entities same software, same URL, zero code changes.
One platform. No fixed mould. That's Zephlo.
Building @sonaroid an energy intelligence platform with AI search over a private indexed database of reports, news, data tables, and infographics.
The unglamorous work: making sure the index is actually right before you let the model anywhere near users.
We ran an audit on our AI chat service this week.
555 out of 556 infographics: never indexed.
Data tables: 0 chunks. Pipeline was never wired up.
The AI was confidently searching through a table that was almost empty. 🧵
Shipped a semantic recommendation engine.
No keywords. Content → vector embedding → pgvector.
User reads something → interest_vector updates → cosine search drives the next feed.
All processing happens in the background. Zero impact on response time.
In February 2024, I made a decision to fully commit to a career in tech. A few months later, I found myself turning down a job offer that came with everything people usually call “sense.”
At the time, I was serving, but mentally I was already deep into tech learning, building, and trying to figure out where I truly fit. There wasn’t much clarity, just conviction.
Then life presented a very comfortable alternative. My uncle called to tell me he was relocating and wanted me to take over his role at a medical science company, Megalife Science.
The company is based in Australia, but they have a branch and exporter here in Nigeria. The role wasn’t shaky or uncertain it was structured, stable, and very real.
There were benefits too. An official car. Travel and airway expenses fully covered whenever you’re out on assignment. Basically, all the things that usually make work stressful were already taken care of.
The manager even reached out personally. At that point, it stopped being advice and became an actual decision waiting for an answer.
I sat with it for a while, because saying yes would have been easy to explain to anyone. It ticked all the boxes people use to measure “progress.” But deep down, I knew it wasn’t the path I wanted to walk.
So I said no, not loudly, not dramatically, just honestly because choosing comfort would have meant walking away from the direction I was already building in tech.
I stayed in Abuja. No official car. No covered flights. No guarantees. Just a laptop, uncertainty, and commitment to a path I believe in.
Question: If you were in my position, would you choose comfort, or stay loyal to your long-term vision?