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

I'm Hasitha Varada, a Senior Data Scientist / AI-ML Engineer with 7+ years of experience building and productionizing machine learning and AI systems across banking, retail, and healthcare. I design low-latency scoring and real-time fraud decisioning platforms, turning complex data into actionable business outcomes. I also integrate GenAI capabilities with Claude and Hugging Face, designing guardrails and Retrieval-Augmented Generation (RAG) pipelines to support investigator workflows. I'm passionate about scalable data engineering, governance, and delivering measurable impact. I collaborate with fraud operations, compliance, and product teams to ensure solutions are auditable, compliant, and enterprise-ready, while continuously exploring cutting-edge GenAI and ML techniques to improve efficiency and decision quality.…I'm Hasitha Varada, a Senior Data Scientist / AI-ML Engineer with 7+ years of experience building and productionizing machine learning and AI systems across banking, retail, and healthcare. I design low-latency scoring and real-time fraud decisioning platforms, turning complex data into actionable business outcomes. I also integrate GenAI capabilities with Claude and Hugging Face, designing guardrails and Retrieval-Augmented Generation (RAG) pipelines to support investigator workflows. I'm passionate about scalable data engineering, governance, and delivering measurable impact. I collaborate with fraud operations, compliance, and product teams to ensure solutions are auditable, compliant, and enterprise-ready, while continuously exploring cutting-edge GenAI and ML techniques to improve efficiency and decision quality.

Hasitha Varada

Data Scientist, AI Engineer, AI Developer, +2





I'm Hasitha Varada, a Senior Data Scientist / AI-ML Engineer with 7+ years of experience building and productionizing machine learning and AI systems across banking, retail, and healthcare. I design low-latency scoring and real-time fraud decisioning platforms, turning complex data into actionable business outcomes. I also integrate GenAI capabilities with Claude and Hugging Face, designing guardrails and Retrieval-Augmented Generation (RAG) pipelines to support investigator workflows. I'm passionate about scalable data engineering, governance, and delivering measurable impact. I collaborate with fraud operations, compliance, and product teams to ensure solutions are auditable, compliant, and enterprise-ready, while continuously exploring cutting-edge GenAI and ML techniques to improve efficiency and decision quality.…I'm Hasitha Varada, a Senior Data Scientist / AI-ML Engineer with 7+ years of experience building and productionizing machine learning and AI systems across banking, retail, and healthcare. I design low-latency scoring and real-time fraud decisioning platforms, turning complex data into actionable business outcomes. I also integrate GenAI capabilities with Claude and Hugging Face, designing guardrails and Retrieval-Augmented Generation (RAG) pipelines to support investigator workflows. I'm passionate about scalable data engineering, governance, and delivering measurable impact. I collaborate with fraud operations, compliance, and product teams to ensure solutions are auditable, compliant, and enterprise-ready, while continuously exploring cutting-edge GenAI and ML techniques to improve efficiency and decision quality.

Available to hire

I’m Hasitha Varada, a Senior Data Scientist / AI-ML Engineer with 7+ years of experience building and productionizing machine learning and AI systems across banking, retail, and healthcare. I design low-latency scoring and real-time fraud decisioning platforms, turning complex data into actionable business outcomes. I also integrate GenAI capabilities with Claude and Hugging Face, designing guardrails and Retrieval-Augmented Generation (RAG) pipelines to support investigator workflows.

I’m passionate about scalable data engineering, governance, and delivering measurable impact. I collaborate with fraud operations, compliance, and product teams to ensure solutions are auditable, compliant, and enterprise-ready, while continuously exploring cutting-edge GenAI and ML techniques to improve efficiency and decision quality.

Skills

Experience Level

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Language

English

Fluent

Work Experience

Data Scientist / AI - ML Engineer at First Citizens Bank

August 1, 2024 - Present

Led the design and production rollout of a real-time fraud and financial crime decisioning platform supporting card, ACH, and digital payment transactions across consumer and small-business portfolios. Built batch and streaming data pipelines using Kafka and PySpark to enable real-time inference and offline model training. Developed scalable feature engineering pipelines in Python (Pandas, NumPy, Polars, Dask) for transaction velocity, device consistency, merchant risk, and longitudinal account behavior. Implemented distributed feature aggregation on Spark with Hive-backed Hadoop tables for regulatory auditability. Built interpretable baseline models (logistic regression) and advanced models (XGBoost, LightGBM, CatBoost) for high-risk segments. Integrated LLM-powered GenAI capabilities to assist investigations with contextual summaries; established prompt guardrails and RAG grounding; applied evaluation and human-in-the-loop validation to ensure reliability. Deployed low-latency infere

Data Scientist at Best Buy

June 1, 2022 - July 1, 2024

Owned the development of an omnichannel demand forecasting and inventory optimization platform across store and e-commerce channels for consumer electronics. Built end-to-end data pipelines using SQL on Google BigQuery and PostgreSQL to integrate historical sales, promotions, pricing, seasonality, and store attributes across millions of records. Created scalable feature engineering workflows (Pandas/NumPy) generating lag features, rolling demand trends, promotion flags, price elasticity signals, and region-specific seasonality. Leveraged PySpark for distributed processing to support large-scale SKU forecasting and efficient batch scoring during peak planning. Implemented baseline ARIMA/Prophet models for transparent validation and extended with XGBoost/LightGBM to capture nonlinear effects of promotions, pricing, and holidays. Applied Bayesian methods (PyMC3/ArviZ) for uncertainty bands. Translated forecast outputs into actionable replenishment recommendations with close collaboration

Data Scientist at Mayo Clinic

June 1, 2019 - May 1, 2022

Analyzed medical claims and EHR extracts for ~120K patients/year, engineering longitudinal utilization, lab, diagnosis, and medication features using SQL on Oracle for modeling and operational reporting. Developed patient risk stratification models starting with interpretable baselines (Statsmodels) and selectively adding tree-based models when lift justified. Implemented readmission risk scoring and care-gap detection across six programs, routing scores to workflow queues and validating impact via pilots and A/B tests. Reduced avoidable outreach by ~12% by prioritizing high-risk cohorts. Built provider quality measurement pipelines with risk-adjusted outcomes across ~300 providers. Automated recurring cohort analyses and scorecards using R/ggplot2; published dashboards in Tableau/Power BI for clinical ops. Processed unstructured clinical text with spaCy and NLTK to improve chart-review triage; reduced noisy flags by ~18%. Explored OpenCV-based imaging quality control in lightweight CN

Data Scientist / AI-ML Engineer at First Citizens Bank

August 1, 2024 - Present

Led design and production rollout of a real-time fraud and financial crime decisioning platform supporting card, ACH, and digital payments; built batch and streaming pipelines using Kafka and PySpark; developed scalable feature engineering pipelines; implemented distributed feature aggregation; built interpretable baseline models and advanced fraud models (XGBoost, LightGBM, CatBoost); integrated GenAI capabilities with Claude and Hugging Face Transformers for contextual fraud investigation summaries; established prompt guardrails and RAG pipelines; evaluated outputs with ML governance; deployed low-latency inference services (FastAPI, Docker, Kubernetes); orchestrated retraining and model promotion with Airflow; tracked experiments with MLflow and DVC; delivered dashboards and monitored model performance; achieved ~20% reduction in false positives and ~10% recall improvement; supported hybrid-cloud environments (AWS/Azure).