VIDUTH ISHARA
Data Scientist
Architecting intelligent systems through computer vision, deep learning, and production-ready AI solutions.
MSc Data Science & Analytics
AI-powered sports analytics projects (published research)
Computer vision | Deep learning | Neural networks | End-to-end ML pipelines
Python | Flask | SQL | Power Platform | Azure
About Me
I'm a Data Scientist and AI Engineer specialising in data intelligence and automation, building production-grade ML systems that deliver measurable business outcomes.
With a background spanning software engineering, data science, and cloud-based automation, I've delivered end-to-end AI solutions for global clients across retail, aviation insurance, and enterprise operations.
I'm the founder of Capital-V Solutions, an AI and operational intelligence consultancy helping operations-driven businesses replace manual processes with intelligent, data-driven systems.
Outside enterprise AI, I'm building a computer vision platform for cricket technique analysis using pose estimation and deep learning to provide objective biomechanical feedback for players and coaches. The foundation is two dissertation projects completed at the University of Westminster.
If you're looking for a Data Scientist or AI Engineer who delivers production systems, not just models, or if your business needs AI automation that actually works, let's connect.
Professional Summary
- MSc in Data Science & Analytics (University of Westminster)
- Published research in pose estimation–based sports analytics
- Built full ML pipelines with custom datasets, CNN models, and real-time inference
- Passionate about ML interpretability, fairness in AI, and building ethical ML systems
All Projects
SpinVision
Deep Learning Analysis of Spin Bowling Technique
AI-powered system analysing left-arm spin bowling using deep learning pose estimation and biomechanical feature extraction. Published MSc dissertation providing objective, real-time feedback on posture, alignment, and release mechanics.
The Wicket Keep Coach
AI Analysis of Wicket-Keeping Technique
Deep learning system evaluating wicket-keeping technique using pose estimation and action classification. Published research paper demonstrating temporal modelling and CNN-based technique scoring.
Retail Forecasting Platform
ML-Powered Demand Forecasting
End-to-end ML forecasting platform automating ETL, running predictive models for every SKU/location, and delivering real-time insights. Achieved 35% improvement in forecast accuracy and significant stockout reduction.
Academic Excellence
“Viduth developed an ambitious deep learning system for real-time pose estimation and spin bowling analysis... exceptionally capable, self-motivated, and technically strong.”
Dr. Yongchao Huang
Visiting Lecturer in Computer Science, University of Westminster
Skills & Technologies
AI & Machine Learning
Programming
Tools & Platforms
Let's Build Something Amazing
Open to AI/ML opportunities in the UK. Let's discuss how I can contribute to your team.