Hi, I’m Semere Herruy 👋 Software engineer building AI systems, backend platforms, and building and iterating toward production-ready systems. I focus on practical models that solve real problems—especially in healthcare, agriculture, and financial decision systems—by combining machine learning and robust backend architecture. I’m actively shipping projects, improving systems, and learning how to scale products from MVP → production.

• AI / Machine Learning (applied) — production RAG, classification, explainability
• Backend Architecture — scalable APIs, Micro-service systems
• AI-driven Products — healthcare, agriculture, credit scoring, automation
• Productization & Integrations — payments, notifications, and deployment

Languages & Frameworks

ML & Data

Databases & Storage

Frontend & UX

Infra & Tools

Integrations Payment flows (Chapa / Payflow integration), FCM notifications, geocoding, mapping

AI question-answering system for Teff and Maize tailored to smallholder farmers. Uses a retrieval pipeline + RAG to answer queries from trusted agricultural documents, with multilingual support (Amharic / Tigrigna). Tech: FastAPI backend, React frontend, document retrieval, structured logging, multi-language handling Repo: https://github.com/semereherruy/ai-agri-advisor-rag — run instructions included in the repo.

Production-ready landing page for an international travel and visa consultancy based in Addis Ababa, designed to convert visitors into clients. Highlights: Conversion-focused UI/UX • SEO optimization with structured metadata • Fully responsive mobile-first design • Performance optimization (lazy loading, code splitting) • PWA-ready with offline support Tech: React • TypeScript • Vite • React Helmet Async Live: https://ethio-travel-agency.vercel.app

A modern, responsive e-commerce frontend inspired by real-world retail platforms, built with a focus on UX, scalability, and production-ready architecture. The project simulates a full shopping experience with a clean, modular frontend design and is structured to easily integrate with real backend services.

Highlights: Product browsing, search, and filtering • Dynamic product detail pages • Shopping cart with quantity management and persistent storage • Complete checkout flow with form validation • Mock-data-driven product system with architecture prepared for REST API integration • Accessibility-first design (WCAG compliant, keyboard navigation, screen reader support) • Performance optimization (lazy loading, image optimization, efficient state handling)

Tech: React • Vite • TypeScript • Tailwind CSS Live: https://mini-ecommerce-app-git-main-semeres-projects.vercel.app/

A production-minded platform connecting patients, doctors, and pharmacies with intelligent matching and real-time notifications. Highlights: ML symptom analysis (sentence-transformer + KNN) • Spatial KNN + fallback matching for nearby providers • Real-time notifications (FCM + WebSockets) • Role-based access, pharmacy workflows, and admin dashboards Tech: NestJS • Prisma • PostgreSQL • React • Socket.IO • Python ML service Code available upon request

End-to-end loan scoring system with LightGBM models, explainability (SHAP), ML inference APIs, and vendor dashboards. Features: Batch scoring and monitoring UI • Model explainability and versioning • Dockerized CI/CD and deployment scripts Code available upon request

Lightweight payments integration example demonstrating checkout and payment confirmation flows using Chapa (Ethiopian payment gateway). Designed to be easily embedded in web checkout pipelines and mobile flows. Repo (example): https://github.com/semereherruy/payflow

• Containerized apps (Docker + docker-compose) for reproducible dev environments
• GitHub Actions for CI/CD and tests
• Logging & structured telemetry for ML + API observability
• Model deployment with FastAPI and organized model versioning

• Telegram: https://t.me/Semere17
• LinkedIn: https://www.linkedin.com/in/semere-herruy-1833222a4/

⚡ “If you knew me yesterday, you don't know me today — I grow daily.”