10,000+ kg
Selected project
Crop2Nutri: AI/ML Predictive Analytics
Crop2Nutri focused on the operational side of AI: getting large and messy data into a form where modeling could guide better decisions. The system combined predictive analytics with RL-driven allocation thinking and scalable data handling.
50+
35%
Overview
Predictive analytics and reinforcement learning for more efficient food allocation across institutions.
The underlying problem was not only prediction, but distribution efficiency. Data had to support forecasts, allocation and ongoing operational decisions across multiple institutions.
01
Structured data pipelines to make large-volume inputs usable for predictive modeling and downstream allocation logic.
02
Used logistic regression and NLP components for predictive insight, while incorporating reinforcement learning agents for allocation decisions.
03
Applied econometric forecasting to reason across institutional demand patterns and likely outcomes.
04
Designed the system with scalability in mind so data movement, model logic and operational use could stay aligned.
01
Helped drive 35% waste reduction across the modeled ecosystem.
02
Established a path toward 25% delivery improvement through RL-based allocation planning.
03
Demonstrated the ability to move between ML, analytics engineering and operational optimization in one project.
Detailed notes
What the work demonstrates.
Operational AI
Rather than limiting the project to one predictive task, Crop2Nutri connected forecasting with action. The useful question was not only what might happen, but what should be allocated next.
Data handling
Large-scale inputs required strong pipeline thinking. The ETL structure supported consistent inputs for predictive models, allocation routines and reporting outputs.
Model blend
The project combined classic methods such as logistic regression with NLP workflows and RL-based decision layers, showing comfort with both traditional and modern AI approaches.
Outcome orientation
The results mattered in practical terms: lower waste, stronger allocation logic and a quantifiable path to better delivery performance.