I'm Chimezie Innocent, a frontend engineer building complex web and mobile products. From AI agent platforms to cloud security and e-learning, I turn demanding requirements into working software and help shape design systems that carry a company's vision across its products.
InnocentChimezie
It started with the movies.Technology looked impossibly cool: code terminals, computer languages that I couldn't fathom, impossible gadgets, and someone typing profusely like the fate of the world depended on it. Well, to be fair, in the movies, the world (howbeit America alone) was at stake, lol!
I loved it all, wanted such a blissful role and profession. I thought I was signing up for this. Needless to say, I totally forgot the country I was in.
But they lied
I went to FUTO, a public university in Nigeria to study Information Technology. Surely the hacker montage was somewhere on the syllabus.
What the movies showed and what the university taught were worlds apart. Actually, make that solar systems apart. Three and a half years later: exams passed. Movies watched. Washed practicals. Still waiting for the hacker montage.
A little fun fact
A little of where I come from belongs in this story too.
I'm Igbo by tribe, from the chief corner states of Anambra. Trade, music, and enterprise are practically a second language around us.
There was rhythm long before there was a keyboard.
20196 months Industrial Training
2020Finals and Project
Mum changed the script
Then came the industrial training year. While other students were hunting for placements that paid and also acquire skills, my mum had a different idea.
"Learn the skills you always wanted" she encouraged me, and even paid for it.
Six months of her money, her belief, and a chance to finally make something real, and my Dad drove me to the office every day and oftentimes even after office hours.
I wanted to learn backend engineering though. However, the advise was to first learn frontend (React), backend (Node), then some React Native before choosing a preferred choice. Although, I did learn them all, I wanted to get a solid grip on frontend first.
I raised my hand
2020 came & now back to the university. The department offered five school projects and encouraged students who had the relevant skills to apply. I applied.
There was an interview. Then, somehow, a bigger project was assigned to me: Frontend developer on a Waste Management and Hazard Alert System.
It became my final-year project. A real problem. Real responsibility.
Another fun fact: The department paid me for my finals project. My first money as a software engineer. Suddenly, this whole frontend thing made a lot of sense.
Last fun fact
I am also a technical writer. I have written over 25 published articles for LogRocket, Prismic, and Strapi.
You can also read my personal writings and thoughts here.
A little room to play. Open the cards on the right for interface experiments, motion studies, and ideas that never made it to production.
After school, I got a job as a Frontend Engineer at CloudPlexo.
I kept building. Products, interfaces, side gigs. The backend-specialist plan quietly moved into a drawer.
I still visit mobile and backend when a side project needs them. Curiosity never really respects a job title.
Still building
Professional work. Side gigs. Open source. Personal projects. Each one added another piece to the way I build.
Wendu, AgentSpec, SleekSign. Pulseguard. Different problems, the same urge to turn an idea into something people can actually use.
The movies got a lot wrong. But the part where making things feels cool? They were right about that.
These are some of the projects I've built both professionally and personally.
These are some of the projects I've built both professionally and personally.
UseOru
SleekSign
Chikọta
Codex Meter
Silo
PulseGuard
Vacay
Bible-Study
Wendu
Pulse-analytics
StoryMagic
UseOru
Loading projects…
01UseOru
02SleekSign
03Chikọta
04Codex Meter
05Silo
06PulseGuard
07Vacay
08Bible-Study
09Wendu
10Pulse-analytics
11StoryMagic
01 / 11
UseOru
Oru is a chat-first AI sales desk. Describe who you want to reach; it finds and scores prospects, writes a sequence in your voice, and holds every send until you approve it.
2026
OnboardingIntegrations page
Problem
Oru was born out of necessity. The sales team kept asking whether we knew any AI platforms they could use for outbound, and whether one could live inside an app we had already built. Everything they tried either sent email on its own or buried the important decision behind a dashboard of filters. So I decided to build one.
The agent can find, write, and queue. It cannot skip you.
A chat for outbound
New chat interface
Oru is a chat-first AI sales desk. You describe who should hear from you, in one sentence: a city, a company size, the job that actually feels the problem. The desk finds and enriches a short list of prospects, scores each one for fit, writes a multi-step sequence in your voice, and holds every send until you approve it.
Leads
Application edit modal
Every draft lands as a card. You can approve it, hold it, or rewrite the line that sounds wrong. Hold is a complete answer. Results show up as artifacts beside the chat: prospect lists, sequences, and charts such as replies this week. Saved lists, tasks, pipeline conversion, and a weekly performance view sit on the desk's home screen.
Application modal
Plugging into the tools sales already uses
Integrations page
Sending happens only after approval, through Composio: Gmail, Outlook, Google Calendar, Calendly, HubSpot, Salesforce, Pipedrive, Attio, Apollo, Hunter, LinkedIn, and Slack, plus any MCP server. The model sees two gateway tools and discovers each integration's tools on demand instead of loading them all up front. Integrations have their own page rather than hiding inside settings.
Chat runs on Kimi through OpenRouter. The yes/no checks (fit, intent, duplicates, and approval) go to a separate judge model through OpenRouter's Decisions API, with a fallback, so the writer never grades its own work. New workspaces start empty and only fill from real actions and connected providers. There are no demo leads hiding in production.
Models selection
Stack:Next.js, React, TypeScript, and Tailwind CSS on the front end. Convex holds the schema, live queries, scheduled AI actions, and crons. Better Auth handles Google sign-in, OpenRouter serves the models, and Composio connects the sending and CRM tools. The chat, streaming, artifacts, and motion are built on Whirl by Anterra, under the MIT license; the sales desk on top is Oru.
