Available for Full-Time & Projects

Hi, I’m Kevin👋

  • AI/ML Architect
  • MLOps Lead
  • Senior Data Scientist
  • Machine Learning Engineer

AI/ML Architect with 9+ years’ experience delivering enterprise-grade solutions across healthcare, fintech, and e-commerce, with expertise in machine learning, deep learning, computer vision, NLP, Generative AI, and MLOps on multi-cloud platforms (AWS, GCP, Azure). Achieved 30% cost savings through optimized ML pipelines, 92% diagnostic accuracy in healthcare AI, and 21% revenue uplift in e-commerce. Recognized for architecting multi-cloud AI platforms, operationalizing large-scale workflows, and leading cross-functional teams while driving innovation with Data Mesh, Lakehouse, LLMOps, RAG-based solutions, and Responsible AI practices.

View Case Studies

Skills

Tools & platforms I use across data engineering, analytics, MLOps, and cloud.

Skills

0 items

Experience

A timeline of hands-on work across healthcare, finance, and insurance—building modern data platforms, streaming pipelines, and analytics.

AI/ML Architect
CitiusTech•05/2022 – Present
  • Architected enterprise-scale AI platforms across healthcare, e-commerce, and financial services.
  • Designed cloud-native ML pipelines on AWS SageMaker, GCP Vertex AI, and Azure ML, enabling 40% faster deployments.
  • Led development of Generative AI and RAG systems using LangChain, Pinecone, and Hugging Face Transformers.
  • Optimized inference infrastructure with ONNX, TensorRT, and Triton Server, reducing latency by 35%.
  • Introduced Responsible AI frameworks (Fairlearn, SHAP, AIF360) ensuring compliance with HIPAA/GDPR.
  • Directed MLOps migration to Kubernetes + MLflow + GitHub Actions, achieving 99% CI/CD success rate.
  • Delivered multimodal AI solutions (text, vision, speech) for cross-domain enterprise clients.
  • Mentored cross-functional teams of 10+ engineers and scientists, improving delivery velocity by 25%.
MLOps Lead
Feedzai•05/2020 – 04/2022
  • Led deployment of ML pipelines for FinTech fraud detection and healthcare diagnostics.
  • Implemented Airflow + MLflow pipelines for orchestrated retraining, cutting downtime by 30%.
  • Containerized models with Docker + Kubernetes, enabling horizontal scaling across 50+ nodes.
  • Deployed edge AI systems with ONNX Runtime and NVIDIA Jetson, accelerating inference by 28%.
  • Automated feature stores with Feast + Snowflake, reducing engineering effort by 20 hrs/week.
  • Introduced model explainability dashboards with SHAP/LIME, increasing regulatory approval rates.
  • Partnered with compliance teams on FDA-ready model documentation for healthcare AI applications.
  • Integrated real-time monitoring with Grafana + Prometheus, improving incident response time by 40%.
Senior Data Scientist
Dataiku•01/2018 – 04/2020
  • Built deep learning pipelines (CNNs, LSTMs) for healthcare image diagnostics, achieving 92% sensitivity in trials.
  • Launched fraud risk models for banking transactions, improving detection precision by 17%.
  • Spearheaded recommendation systems for e-commerce, boosting CTR by 21%.
  • Enhanced data pipelines using PySpark + SQL, cutting processing time for TB-scale datasets by 40%.
  • Devised NLP models (BERT, spaCy) for document classification with 89% accuracy.
  • Partnered with clinicians and risk teams to integrate AI models into production decision systems.
  • Conducted A/B testing and uplift analysis, validating model improvements in live environments.
  • Created reproducible ML workflows using DVC + Git, improving collaboration across teams.
Machine Learning Engineer
Uniphore•01/2016 – 12/2017
  • Formulated ML models using scikit-learn, XGBoost, logistic regression for early e-commerce analytics.
  • Systematized feature engineering pipelines with Python + Pandas, reducing manual effort by 50%.
  • Deployed recommendation prototypes with Flask APIs, improving product discovery for users.
  • Implemented time-series forecasting (ARIMA) for retail demand prediction, improving accuracy by 15%.
  • Built fraud scoring models with decision trees + logistic regression, reducing false negatives by 12%.
  • Collaborated with engineers to move ML models from notebooks into production environments.
  • Engineered data ingestion workflows in SQL + Airflow, supporting millions of transactions daily.

Projects

Selected builds across healthcare, finance, and insurance—real-time pipelines, modern data platforms, and analytics that ship value.

Real-Time Fraud Risk Platform (FinTech)

  • Developed fraud detection & credit scoring engine integrated into transactional APIs using XGBoost, LightGBM, SHAP, Kafka, and Spark.
  • Increased fraud capture by 45%, boosting compliance readiness and cutting false positives by 18%.

Clinical AI Imaging Platform (Healthcare)

  • Designed CNN-based diagnostic pipeline for X-ray and OCT scans using TensorFlow, PyTorch, and OpenCV, achieving 92% diagnostic accuracy in HIPAA-compliant deployments.
  • Enabled real-time inference on clinical edge devices with ONNX integration, ensuring 92% diagnostic sensitivity.

Intelligent Retail Forecasting Engine (Retail)

  • Built hybrid LSTM + Prophet model for SKU-level demand forecasting with scalable MLOps pipelines (Airflow, MLflow, GCP Vertex AI).
  • Generated 33% higher forecast accuracy and reduced retail stockouts by 22%.

Enterprise Knowledge Assistant (Generative AI)

  • Architected RAG-based enterprise knowledge assistant using LangChain, Pinecone, Hugging Face Transformers, and AWS SageMaker.
  • Reduced manual research time by 40% and improved query accuracy to 93% across multi-domain datasets.