Looks like you have JavaScript disabled. For the full Twine experience, you will need to re-enable it.

I'm W.P. Roshan, a Data Science Engineer with 11+ years in the AI/ML industry. I specialize in building and deploying ML and DL models, including LLMs and RAG systems, and I love turning complex data into actionable insights. I enjoy solving real-world problems across Healthcare, Digital marketing. Telco, Insurance, and EduTech with scalable AI/ML solutions. I'm proficient in Python and MLOps, including Docker and Kubernetes, and I focus on optimizing ML pipelines, governance, and scalable AI solutions. I'm excited to bring my experience to AI-driven innovation.…I'm W.P. Roshan, a Data Science Engineer with 11+ years in the AI/ML industry. I specialize in building and deploying ML and DL models, including LLMs and RAG systems, and I love turning complex data into actionable insights. I enjoy solving real-world problems across Healthcare, Digital marketing. Telco, Insurance, and EduTech with scalable AI/ML solutions. I'm proficient in Python and MLOps, including Docker and Kubernetes, and I focus on optimizing ML pipelines, governance, and scalable AI solutions. I'm excited to bring my experience to AI-driven innovation.

W.P. Roshan

Data Scientist, AI Engineer, Data Analyst, +1





I'm W.P. Roshan, a Data Science Engineer with 11+ years in the AI/ML industry. I specialize in building and deploying ML and DL models, including LLMs and RAG systems, and I love turning complex data into actionable insights. I enjoy solving real-world problems across Healthcare, Digital marketing. Telco, Insurance, and EduTech with scalable AI/ML solutions. I'm proficient in Python and MLOps, including Docker and Kubernetes, and I focus on optimizing ML pipelines, governance, and scalable AI solutions. I'm excited to bring my experience to AI-driven innovation.…I'm W.P. Roshan, a Data Science Engineer with 11+ years in the AI/ML industry. I specialize in building and deploying ML and DL models, including LLMs and RAG systems, and I love turning complex data into actionable insights. I enjoy solving real-world problems across Healthcare, Digital marketing. Telco, Insurance, and EduTech with scalable AI/ML solutions. I'm proficient in Python and MLOps, including Docker and Kubernetes, and I focus on optimizing ML pipelines, governance, and scalable AI solutions. I'm excited to bring my experience to AI-driven innovation.

Available to hire

I’m W.P. Roshan, a Data Science Engineer with 11+ years in the AI/ML industry. I specialize in building and deploying ML and DL models, including LLMs and RAG systems, and I love turning complex data into actionable insights. I enjoy solving real-world problems across Healthcare, Digital marketing. Telco, Insurance, and EduTech with scalable AI/ML solutions.

I’m proficient in Python and MLOps, including Docker and Kubernetes, and I focus on optimizing ML pipelines, governance, and scalable AI solutions. I’m excited to bring my experience to AI-driven innovation.

Skills

Experience Level

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Intermediate

Intermediate

Language

English

Fluent

Sinhala, Sinhalese

Advanced

Work Experience

Lead Data Science Engineer at Andela

January 1, 2024 - November 27, 2025

Led the design and deployment of scalable ML systems, including anomaly/outlier, churn, and automatic valuation models on serverless platforms. Built multi-modal, agentic RAG pipelines and CDP integrations, and implemented ML observability, data lineage, and model monitoring using ZenML, MLflow, BentoML, and CometML. Replaced legacy pipelines with modern FTI-based architecture, boosting usability and performance by ~40%. Implemented data versioning and governance via DVC, Great Expectations, and Apache Atlas, achieving higher data quality and governance. Leveraged LangChain and LlamaIndex to develop LLM-backed AI apps.

Senior Data Science Engineer at Gapstars (Client: Spotr.ai, Netherlands)

December 1, 2023 - December 1, 2023

Built a remote property inspection AI system processing 500+ properties monthly. Developed CV-based risk assessment and underwriting models and automated damage assessment for claims using Mask R-CNN and Faster R-CNN. Implemented roof material classification using CNNs with LiDAR data, and deployed remote sensing workflows. Reengineered Airflow pipelines with mage.ai, increasing performance by ~60%, and strengthened data governance with Great Expectations and Apache Atlas by ~75%. Introduced LLM- and GPT-enhanced agentic RAG pipelines for real estate and finance sectors.

Senior Data Science Engineer at Adventus Education

May 1, 2023 - May 1, 2023

Architected big data pipelines and ETL processes to improve EdTech platform performance by 40%. Built ML models to identify at-risk students with ~92% accuracy and drove a 23% reduction in at-risk student levels. Enhanced churn prediction and retention strategies with ~93% accuracy.

Lead Data & BI Engineer at DataMTX Labs

March 1, 2022 - March 1, 2022

Led a high-impact data engineering team, delivering projects 90% faster and achieving 60% improvement in data processing efficiency. Implemented data warehouse (Redshift) and data lake (S3), driving a 70% efficiency gain. Built time-series forecasting models (ARIMA, Auto-ARIMA, SARIMA, SARIMAX) with 89–94% accuracy, improving forecast precision by 40%.

Senior Data Science Engineer at DigitalXLabs

October 1, 2021 - October 1, 2021

Architected AdStudio platform for AdTech using Spark, Kafka, Druid, and Cassandra, increasing efficiency and scalability by ~30%. Developed price optimization and recommendation models using CDP, DMP, and CRM data with 97% accuracy, boosting conversions by 7%.

Big Data Team Lead at Cloud Solutions International

July 1, 2019 - July 1, 2019

Automated data workflows with Apache Airflow, improving data accuracy by 30% and reducing manual errors. Enhanced IoT metrics visibility by 50% via ELK stack, enabling faster decision-making.

Senior Engineer (AI/ML) at IQVIA (formerly IMS Health)

January 1, 2019 - January 1, 2019

Built real-time drug adverse event tracking using Spark, Hadoop, Kafka, Nifi, and Nutch; created KPI dashboards improving response times by ~40%. Developed OCR/HTR services with PyTorch (CTC) achieving ~85% transcription accuracy and 20% faster processing. Implemented NLP/NLU systems with CoreNLP, SparkNLP, spaCy, and NLTK achieving strong F1 and entity recognition.