Maiwand Jan Alamzoi

ZaminAI — Satellite Agricultural Intelligence

Designed and deployed a full-stack AI platform solo, turning live Sentinel-2 satellite data into field-level NDVI, water-stress and crop analytics. Built a multilingual AI advisory assistant (Claude / Gemini) answering farming questions in English, Dari and Pashto. Live MVP, free and publicly explorable at zaminai.org, built to close the digital-agriculture gap for the 500M+ smallholder farmers worldwide who currently lack this kind of support. Collaboration with Wageningen University & Research.

Result: live, publicly explorable MVP at zaminai.org, built and deployed solo — aimed at closing the digital-agriculture information gap for 500M+ underserved smallholder farmers.

Crop Stress Prediction — Afghanistan / Netherlands / New Zealand

Open-source pipeline predicting vegetation-stress risk 2–4 weeks ahead from Sentinel-2 imagery and public weather data, across 36 farm locations in three countries with very different agriculture. Random Forest and XGBoost, evaluated on a genuine 2023–2024 holdout (F1 0.41, 72% accuracy — real numbers, not rounded up), with SHAP explainability for every individual prediction and a live interactive dashboard.

Netherlands Climate Risk & Nitrogen Dashboard

Open-source dashboard covering flood, drought, heat, land-subsidence and nitrogen-deposition risk for 34 locations across the Netherlands, built entirely on official Dutch government data (Climate Impact Atlas, RIVM GCN/GDN) — no modeling, exactly what the government's own services report, with the same categories and units as their official legends.

Result: Nijmegen shows the highest nitrogen deposition of all 34 locations, and agriculture (NHx) outweighs traffic (NOy) everywhere — matching the real Dutch nitrogen-crisis picture.

Netherlands Scope 1+2 Emissions Calculator

Open-source Scope 1+2 GHG calculator built on official Dutch emission factors (co2emissiefactoren.nl — the standard reference for CO2-Prestatieladder/CSRD reporting), not a generic international substitute. Enter real fuel and electricity data, get a real emissions estimate, every number traceable to its official source.

Result: verified real calculation — 1,000 litres of petrol + 500 kWh grey electricity = 3,300.5 kg CO2e, checked by hand against the app's output.

Crop Stress & Disease Detection — Agrisim & Rapagra

Built and validated ML models (Random Forest, XGBoost) on multispectral, meteorological and field data for early stress and disease-risk detection, reaching ~87% classification accuracy. Deployed a monitoring application used across multiple commercial farm sites, part of EU-funded innovation projects in Wageningen.

Result: automated monitoring across multiple commercial farm sites with ~87% prediction accuracy, cutting manual field-checking time.

Sensor-Driven Monitoring — Circle Farming

Developed IoT-based monitoring and real-time data workflows for controlled-environment (circular) farming, integrating live sensor streams into actionable insights and closed-loop optimisation of growing conditions.

FAO On-Farm Water Management — Afghanistan

As FAO (UN) National Consultant, delivered food-security assessments and policy analysis; built a national database of agriculture and irrigation policies, strategies and laws. Coordinated FAO, government ministries and development partners. Selected onto the FAO UN Expert Roster.

Afghanistan Development Initiative — Open Data & Research Platform

Founded and built a free, open-access research and data platform for Afghanistan's development, spanning six pillars for practitioners, researchers and policymakers worldwide. Published policy briefs on water security, soil degradation and economic diversification, plus a nutrient-use-efficiency research note.

Result: live, publicly explorable platform with 4 published policy briefs/research notes, built and maintained solo.

National Mechanization Services — Ministry of Agriculture (MAIL)

Led a team of 15+ delivering agricultural mechanization services across all 34 Afghan provinces — machinery deployment, field-testing and redistribution, and training farm managers and operators on machinery and technology, including conservation and zero-tillage practices.

IFA × Wageningen University & Research Hackathon — Winner

Won the 2026 hackathon on Nutrient Use Efficiency Tracking, applying data-driven modelling to real-world agronomic decision-making challenges.