OTseun.wtf
Currently listening

Nothing playing right now

Start a track and this panel will update.

01 / Selected work

Things I've shipped.

01May 2025 — Present

Trex Workspace

Software Engineer

Core platform APIs for projects, agile, and collaboration—chat, notifications, appointments, and settings—with WebSocket updates, background workers, and Gemini-backed document/PDF AI.

APIsWebSocketsWorkersGemini
03Mar — May 2025

Farm Monitor Africa

AI Engineer

Adaptive farming calendars, GAP advice, and weekly yield prediction from soil, weather, NDVI, and moisture indexes.

OpenAIWeather dataNDVI

Featured case study · AI Engineer

Helping farmers see the season ahead.

At Farm Monitor Africa, I built the AI systems that turn soil, weather, and crop signals into practical guidance—assistant, calendar, GAP advice, and weekly yield prediction.

Case 001 · Farm Monitor Africa · Mar — May 2025

farm.monitor / africalive

Plot · Cassava · Ogun

Week 14 · flowering stage

GAP ready

NDVI

0.71

Soil H₂O

34%

14-day

Rain+

AI assistant

Delay fertilizer until Thursday—soil moisture rising with the forecasted front.

MonCalendar regenerated · weather shift
WedGAP tip · weed pressure window
FriYield model · weekly forecast ready

01 · Challenge

Farm decisions were scattered across weather, soil, and season.

Farmers using Farm Monitor Africa needed more than a dashboard—they needed guidance that could keep up as conditions changed. Static schedules and generic advice could not account for soil data, forecasts, and crop stage together.

02 · Intelligence layer

An AI assistant grounded in farming practice.

I designed and built an OpenAI-powered agricultural assistant that turns farm context into clear recommendations, plus a Good Agricultural Practices (GAP) engine for personalized advice on what to do next—tied to real-time weather.

03 · Adaptive calendar

Schedules that regenerate when the weather shifts.

I built a farming calendar generator that adapts to soil data, forecasts, and seasonal change, then wired auto-triggered regeneration when weather fluctuates—so crop plans stay honest instead of going stale.

04 · Prediction

Weekly yield foresight from field signals.

I developed a weekly yield prediction system using historical data, 14-day forecasts, NDVI, and soil moisture indexes—giving farmers an earlier read on how the season is tracking.

OpenAIAI engineeringWeather dataNDVIGAPYield models

02 / A little context

Systems first. Clarity always.

I'm a backend and AI engineer—and co-founder of Turinstack, a design studio where product, systems, and interfaces ship together.

I care about

Reliable APIs, workers that keep HTTP thin, honest data models, and AI that earns its place with real outcomes—not demos.

Stack, lately

TypeScript / Golang / Python
Postgres / Kafka / Redis