A Product-Focused Engineer
I build real products, not demos.
Most of my work lives at the intersection of full-stack engineering, AI systems, and product thinking — turning messy, manual workflows into clean, automated software people actually use.
I’m currently building quietly.
- Shipping end-to-end products (idea → system → users)
- Designing AI systems with clear guardrails and real utility
- Automating boring, error-prone processes
- Building tools for education, operations, and internal teams
- Thinking like a founder, even when writing client's code
I don’t chase trends.
I chase problems that shouldn’t exist anymore.
An automated system for grading handwritten theory exams, built to reduce bias and inconsistency in manual marking.
- Visual Language Models interpret handwriting & diagrams
- LLM agents grade against official marking guides
- Async job processing + real-time progress updates
- Designed for scale and fairness in large exam systems
Stack: Next.js, FastAPI, LangGraph, PostgreSQL, Redis, RabbitMQ, Docker
An AI-powered educational platform that generates structured, illustrated learning articles from any topic.
- Custom prompt pipelines (no heavy AI frameworks)
- Image + content generation orchestration
- Built serverless, clean, and fast
Stack: Next.js, TypeScript, prisma, OpenRouter
An AI tool used by 50+ students to generate realistic internship logbooks in minutes.
- Smart activity splitting
- Multiple writing styles (technical, concise, detailed)
- Built from a personal pain point → real adoption
Stack: Next.js, TypeScript, OpenRouter
A domain-specific RAG-based assistant for hospitals.
- Vector-grounded responses
- Carefully engineered system persona
- Manually deployed on AWS EC2 (Linux, NGINX, SSH)
Stack: Next.js, OpenRouter, Vector DB, AWS EC2
Built multiple production systems for schools:
- Bursary Management Dashboard (finance, invoicing, analytics)
- School Website & CMS
- Chapel Letter Automation System (desktop automation)
Focused on replacing manual institutional workflows with software.
Frontend
- Next.js (App Router)
- Tailwind CSS
- ShadCN UI
Backend
- Convex
- Better-Auth
- Server Actions
- express
- fast api
AI
- Vercel AI SDK
- LLM / VLM orchestration
- RAG pipelines
Infra
- Vercel
- Modern, serverless-first deployments
I think in systems, not features.
Every project asks:
- What’s the real bottleneck?
- What breaks at scale?
- What should never be manual again?
If a solution isn’t usable, maintainable, and explainable — it’s not done.
- LinkedIn: https://www.linkedin.com/in/daniellucasadebayo
- X: https://x.com/coollad49
- Email: daniellucasadebayo@gmail.com
Quietly building. Shipping deliberately.