Hi, I’m
Sahan Dissanayaka
Data Scientist II (MLOps) at ISA, Air Arabia Group, working on accelerated data science,real-time inference systemspropensity modelling, and reinforcement-learning-based decision systems — with a research background in graph neural networks from my PhD fellowship at UCF.
Turning data into decisions
A quick look at who I am and what drives my work.
I’m a Data Scientist II (MLOps) at ISA, Air Arabia Group, where my work centers on three streams: end-to-end workflow design with MLOps, propensity modelling for customer behaviour, and reinforcement-learning-based decision systems for consumers — built and shipped through production MLOps pipelines.
I have a strong research background PhD fellowship in Industrial Engineering at the University of Central Florida, researching time series, anomaly detection, eometric deep learning, graph neural networks and transformer architectures for drug discovery and predictive maintainance, before returning to industry. That research still shapes how I approach applied ML problems today.
My toolkit spans Time Series, Machine Learning, MLOps, network science, and Reinforcement Learning, and I’m driven by turning ambitious ideas into systems that work reliably in production. I’m currently open to new opportunities in applied ML and MLOps/LLMOps.
Skills & Technologies
The stack I use to research, build, and ship machine learning systems.
Data & Time Series
Machine Learning & Deep Learning
Reinforcement Learning & Research
MLOps & Cloud
Agentic AI & LLMs
Data Engineering & Responsible AI
Projects & Research
Applied MLOps work at ISA, Air Arabia Group, alongside open-source and PhD research.
Time-Series Forecasting & Anomaly Detection
Forecasting demand and revenue signals and detecting anomalies in operational time series at ISA, plus an open-source package for benchmarking time-series forecast error metrics — published to PyPI and npm.
View on GitHubPropensity Modelling for Customer Behaviour
Building propensity models that predict customer behaviour — purchase likelihood, upsell, and engagement — to power targeted, revenue-driving decisions across the airline group.
Read moreReinforcement Learning for Decision Systems
Developing RL-based models for sequential decision-making — dynamic pricing and resource allocation — and operationalising them through production MLOps pipelines.
Read moreMLOps Pipelines on Azure
Built and deployed end-to-end ML pipelines — automated training, validation, and deployment — for revenue-optimization models, and demoed the full Azure MLOps workflow at Intellihack 3.0.
View on GitHubMolecular Property Prediction with GNNs
PhD research at UCF: graph neural network architectures that learn from 2D and 3D molecular conformations to predict quantum, chemical, and ADME properties — supporting interpretable drug discovery and molecular generation.
Read moreNetwork Science Research
Modeling complex interaction networks — from drug-protein interactions to social influence — using graph theory and machine learning to surface anomalies and high-impact nodes.
View on GitHubExperience & Education
From undergraduate research to a PhD fellowship and back to applied ML in industry.
Data Scientist II, MLOps
ISA, Air Arabia GroupLeading work on time-series forecasting, propensity modelling, and reinforcement-learning-based decision systems — designing, training, and deploying models through production MLOps pipelines.
PhD Fellowship, Industrial Engineering
University of Central FloridaResearched geometric deep learning for molecular property prediction, drug discovery, and molecular generation using graph neural networks and transformer models.
Data Scientist
Air ArabiaApplied machine learning to revenue optimization and built MLOps pipelines on Azure for automated model training, validation, and deployment.
GSoC Contributor & Mentor
SCoRe LabContributed to open-source machine learning tooling during Google Summer of Code 2021, and returned as a mentor to guide the next cohort of contributors.
BSc (Hons) Computer Science
University of Colombo School of ComputingGraduated with a CGPA of 3.84/4.00, with a focus on machine learning, network science, and software engineering.
Let’s build something great
Open to new opportunities in applied ML & MLOps, research collaborations, and conversations about time series, propensity modelling, and reinforcement learning. Drop me a line.
Say Hello