DO.
CV

Software Engineer | Mobile · Desktop · AI

Daniel
Oluremi

I don’t just write code - I build products. I’ve worked on startup teams across fintech, social, and agritech - building mobile apps, desktop apps and training AI models - and that experience taught me to think beyond the task in front of me to what the market wants, what a client is really asking for, and which technical decisions will still make sense several months/years down the line. Bring me in to build AND to help decide what’s worth building and how best to do it.

Apps worked on
10+
Shipped to production
4 + 1
Downloads across stores
7k+
Mobile dev experience
4 years, and counting
Open-source projects
5 - four plugins, one dataset
Articles written
1
Degree
B.Sc. CS, first-class - best in department
Currently
AgeOfAI (SumeruEdutech)

Selected work

Farm dashboard with weather, season summary, and a satellite view of the field Farm fields list with search and add-new-farm Adding a new farm - satellite mapping with location details Satellite crop-health imagery with a health index scale Farm activity log with dated entries and costs

Farm Monitor

Mobile App Developer · 2024

A remote agro project management and monitoring platform built for Africa - real-time farm visibility for farmers, financiers, and stakeholders. I built farmer-side features in the Flutter app.

  • Offline-first field data - a sync layer of ~20 dedicated syncers over SQLite records harvests, sales, labour, and logistics without connectivity
  • Satellite field mapping with Google Maps and crop-health monitoring dashboards
  • Localized in English, Hausa, Igbo, and Yoruba, with face-verification onboarding

1,000+ downloads · 4.3★ on Google Play

FlutterSQLiteGoogle MapsProviderSentry

DripoolPay bills purchase home screen with wallet balance Airtime top-up with discount pricing Referral earnings screen Transaction history with downloadable receipts Wallet top-up via bank transfer

DripoolPay

Built solo · 2024

A bills-and-wallet fintech app - airtime, data, subscriptions, meter payments. The one I built entirely alone: every screen, pixel-perfect to its UI design, hardened for real money.

  • Custom PIN keypad, biometric unlock, and encrypted secure storage
  • Transaction alerts via FCM with local notification fallbacks
  • 15 releases shipped to production

4.7★ on Google Play

FlutterRiverpodlocal_authSecure storageFCM

RecenthPost onboarding - 24-hour categorized posts Feed screen making connections across the globe Daily challenges screen User profile with rated reactions Realtime chat conversation

RecenthPost

Mobile App Developer at Bakkaz · 2024

A 24-hour categorized social platform where every post expires and the timeline stays current by design. I built much of the Flutter app - feeds, live streaming, realtime chat, and voice notes.

  • Live-streaming module plus Socket.IO chat running side by side
  • Voice messages end to end: recording, waveform rendering, playback
  • Offline caching with SQLite and biometric app lock

1,000+ downloads · 5.0★ on Google Play

FlutterLive streamingSocket.IOflutter_soundSQLite

Errandaar home screen with errand options and top-rated errandees Finding trusted errandees near you Tracking an errand’s progress in real time on a map Connecting with verified errandees Errandaar wallet with secure payment options

Errandaar

Mobile App Developer at Bakkaz · 2025

A two-sided errand marketplace - people who post errands and runners who complete them. I built core parts of the Flutter client: one codebase, two personas, realtime everything in between.

  • Realtime errand posting, chat, and order tracking over Socket.IO and raw websockets, alongside FCM push
  • Client-side AWS Signature V4 request signing for direct media uploads
  • Face-verification KYC (Regula SDK) and Google Maps live location through 37 shipped builds

5,000+ downloads · live on both stores

FlutterSocket.IOGoogle MapsAWS SigV4FCM

Hands holding a phone running Balancc’s Obligations screen - monthly obligations, targets, and spending categories

Balancc

Mobile App Developer · 2025 - 2026

A financial-insights app that connects to bank accounts, categorizes transactions, and turns raw spending into decisions. On a small team, I built significant parts of the Flutter app, spanning 13 feature modules across accounts, budgets, transactions, and insights.

  • Background sync pipeline - WorkManager plus foreground services keep accounts fresh without draining the battery
  • Home-screen widgets and a dual charting stack (Syncfusion + fl_chart) for the Insights module
  • Shorebird over-the-air patching, biometric lock, and jailbreak detection in production

FlutterRiverpodWorkManagerFirebaseShorebird OTA

In the works - at Eleven36, since Dec 2025

Cloutt

A creator platform nearing launch: live streaming with AR face filters - DeepAR piped into Agora RTC through a custom bridge - a virtual-gift economy with wallet top-ups and withdrawals, realtime chat and presence over Socket.IO, and a full in-app video editing pipeline. I build core features across its live-streaming and monetization stack.

FlutterAgora RTCDeepARSocket.IOFirebase

SecuRent

A rental platform for the Nigerian market: tenant vetting and KYC, rent payments and disputes, PDF rent receipts - and landlord-tenant messaging end-to-end encrypted with the Signal Protocol, on custom Hive-backed key stores. I build core features across its vetting, payments, and encrypted-messaging stack.

FlutterSignal ProtocolPaystackRiverpodHive

AI engineering

Models, datasets, and the infrastructure under them.

Real machine-learning work, not API wrappers: models fine-tuned and evaluated properly, a dataset published for others to use, and the engineering that makes training repeatable.

Teaching a 1.5B model step-by-step math

Built a tutoring dataset from NCERT textbooks (PyMuPDF extraction) merged with Orca-Math and MetaMathQA, then QLoRA fine-tuned Qwen2.5-Math-1.5B-Instruct - 4-bit, LoRA r=32 across attention and MLP projections. GSM8K and MATH were excluded from training and used only as held-out benchmarks - a baseline-vs-finetuned eval on GSM8K and a held-out eval on MATH.

PyTorchUnslothQLoRATRL/PEFTGSM8K evals

A Yoruba dataset the community can use

Scraped and cleaned Yoruba Wikipedia through the MediaWiki API into an academic QA dataset, published it on Hugging Face, and fine-tuned Qwen2.5-1.5B on it - low-resource-language NLP for a language 40+ million people speak.

MediaWiki APIDatasetsQLoRAHugging Face

One family, six specialist models

SchoolExl runs on Sumeru SLMs - six small language models distilled for education: lightweight everyday assistants that run even offline, step-by-step thinking and reasoning models, a voice model that listens and speaks, and a vision model that reads pages, diagrams, and handwriting.

DockeruvHugging Face stackCIRunPod

Open source

Plugins other builders can use.

Four Flutter plugins on pub.dev, born from the AR live-streaming work in Cloutt - camera to AR to RTC, packaged.

Archive

Also built.

Experience

Where I’ve worked.

  1. Aug 2025 - now

    AgeOfAI (SumeruEdutech)

    Mobile/Desktop Engineer and AI Engineer - the SchoolExl mobile app, its offline EdgeAI desktop platform, and training the Sumeru SLMs.

  2. May 2025 - Jan 2026

    Balancc

    Mobile App Developer - financial-insights app; background syncing, charts, home-screen widgets, OTA patching.

  3. May 2024 - Jun 2025

    Bakkaz Information & Technology

    Mobile App Developer - shipped RecenthPost and Errandaar to the stores, faithful to their designs.

  4. Feb - Jun 2024

    Farm Monitor

    Mobile App Developer - farmer-facing app for an agricultural finance and monitoring platform.

Portrait of Daniel Oluremi in a white shirt against amber curtains

About

Beyond the code.

I graduated top of my Computer Science class at the Federal University of Agriculture, Abeokuta - first-class honours, best graduating student in the department. I’ve been writing programs since 2019 and building mobile apps since 2022 - Flutter across mobile and desktop - and lately machine learning: from DeepLearning.AI’s specialization into fine-tuning real models on real GPUs.

Most of my work has happened on startup teams - designers, backend engineers, founders - where the job was never only the code. It was understanding what the market wanted, what the client actually meant, and which path would get a product noticed. I’ve watched approaches fail and learned the ones that hold up. I’m a naturally observant person, and it shows in what I build: I pay attention to the often-overlooked details that leave the strongest impressions on users.

Contact

Let’s build something
people rely on.

Open to full-time roles and select freelance work - remote-friendly.