Thamindu
Gunasinghe
Transport and Logistics Engineering undergraduate applying AI, Machine Learning and Explainable AI to real-world supply-chain forecasting, procurement, and inventory-management problems.
Recent work includes a coconut price forecasting model at Nestlé that achieved over 95% accuracy, supporting a procurement area with roughly LKR 17B in coconut-related spend within an estimated LKR 60B budget.
Where logistics engineering meets machine learning.
I'm a Transport and Logistics Engineering undergraduate at the University of Moratuwa, focused on making AI genuinely useful for supply chains — forecasting, procurement and inventory decisions that people can actually trust.
My final-year research is on Explainable AI for Demand Forecasting and Inventory Management Systems — combining forecasting accuracy with business interpretability, so supply-chain users understand demand drivers, inventory risks, and recommended actions.
Experience
2024 → PresentNestlé Lanka PLC
- Initiated and developed a coconut price forecasting project supporting procurement planning and budget visibility for a strategically important raw material.
- Built a machine-learning model achieving over 95% forecasting accuracy during evaluation, improving price visibility for procurement decisions.
- Worked within a procurement area of roughly LKR 60B, with coconut-related spending contributing around LKR 17B.
WIWIS.AI
- Managed delivery of AI-powered projects: Mr. Wee (AI WhatsApp sales assistant), Bus Passenger Tracking & Revenue Assurance, and the PrimeCare healthcare logistics platform.
- Coordinated cross-functional teams, bridging transport and logistics expertise with AI/ML engineering workflows.
- Applied prompt engineering and AI-assisted development to speed up design, prototyping and deployment.
Projects & Research
09 — Full recordExplainable AI for Demand Forecasting
Interpretable demand-forecasting and inventory-management system that explains model outputs — combining accuracy with explainability so users understand demand drivers, inventory risks and recommended actions.
Coconut Price Forecasting
Initiated and developed a forecasting model for coconut price prediction, a crucial procurement input. Achieved over 95% accuracy in evaluation, supporting budget planning for ~LKR 17B in spend.
Bus Passenger Tracking
GPS and route-data analytics for revenue assurance, plus a route planner web app feeding route sections and fare details into the system.
Sopaka.AI
AI assistant for cleaning and standardizing supply-chain datasets — automating the messy first mile of every analytics project.
Driver Drowsiness & Phone Use Detection
AI-based monitoring to improve driver alertness and road safety, detecting drowsiness and phone use in real time.
PrimeCare Logistics Platform
Project management for a healthcare sample delivery ecosystem — web and mobile systems for coordination and dispatch.
Automated Storage & Retrieval System
Built the full dashboard in Next.js and microcontroller control in Arduino for an ASRS prototype — inventory tracking and pick/place workflow end to end.
Intelligent Traffic Light System
Research applying computer vision and AI for traffic-flow optimization at signalized intersections.
Mr. Wee — RAG Prototype
Managed the prototype of an AI-driven WhatsApp commerce assistant built on LLMs and RAG pipelines.
Skills
Applied in real-world projects- Python
- PySpark
- TensorFlow
- FastAPI
- Scikit-Learn
- OpenCV
- ONNX
- Next.js
- C#
- OpenAI API
- LLaMA
- Qwen
- RAG pipelines
- Prompt engineering
- Azure Functions
- Logic Apps
- MongoDB
- Runpod
- PostgreSQL
- Google Cloud
- Docker
- Kubernetes
- Systems engineering
- Operations engineering
- Operations research
- Supply chain engineering
- FreeCAD
- SUMO
- AnyLogic simulation
- Canva
- Photoshop
I collaborate effectively with AI and software engineers. Through AI-assisted development and prompt engineering, I apply these technologies in real-world projects.