Lead Data & AI Engineer

I lead the design and deployment of an Agentic AI framework on top of Kafka and Flink streams for maritime operations.

My work includes RAG over internal technical documentation, embedded ML inference inside Flink operators, and production monitoring for model drift, latency, and quality. I also help define evaluation protocols and observability practices for LLM-based systems.

I act as a technical lead within a larger program, coordinating with data scientists, business experts, and platform teams to deliver reliable AI and data products.

I built and operated a real-time streaming platform for fleet-wide navigation, engine, and IoT data.

My responsibilities included Apache Flink stateful jobs, Kafka backbone design, Iceberg lakehouse sinks, and cloud migration to AWS. I also contributed to predictive maintenance and early GenAI/RAG use cases built on the same streaming foundation.

This role focused on low-latency processing, high availability, and scalable data delivery for downstream consumers and analytics systems.

I migrated monolithic SQL batch workloads into distributed streaming and batch architectures.

I delivered the first Apache Flink jobs in production, built Spark pipelines on AWS EMR and Hadoop-YARN, and helped establish the Kafka and streaming foundations for later real-time initiatives.

I also worked on feature engineering at scale and cloud cost optimization through more efficient processing patterns and infrastructure choices.

I developed and deployed predictive maintenance solutions for wind and solar farms using sensor telemetry.

My work covered data pipelines, feature engineering, and machine learning models for failure prediction and operational optimization. I also applied constrained mathematical optimization to support decision-making.

This role combined cloud data engineering, applied machine learning, and industrial problem solving.

I built supervised churn and pricing models for auto insurance contracts.

I worked on dynamic pricing, elasticity analysis, explainability with SHAP, and end-to-end ML pipelines in PySpark and Databricks. I also used MLflow for tracking, versioning, and monitoring.

The work had direct business impact through revenue improvement and better actuarial decision support.

I worked on the full lifecycle of computer vision models, from data labeling to training and deployment.

I built Flask REST APIs for real-time classification and contributed to internal AutoML experimentation. I also supported data integration with GCP services such as BigQuery and Firebase.

This internship gave me early hands-on experience in ML deployment and production-oriented engineering.

I studied machine learning, deep learning, constrained optimization, stochastic modeling, data analysis, and programming.

The program also covered SQL, NoSQL databases, and software development in Python, Java, and C++.