Senior AI Engineer

I lead the AI team and set technical direction for agentic orchestration, tool calling, and system prompting. I align technical standards to enable LLMs to answer complex questions grounded in technical documents, time-series data, and P&ID diagrams.

I designed a Mastra-based question-answering workflow engine with multi-hop retrieval, dependency-aware reasoning, and parallel fan-out across tools and knowledge sources. I also coordinated engineering, product, and domain teams around a unified knowledge graph spanning more than 200 mixed files.

I partnered with security and compliance stakeholders to deliver data sanitization pipelines with adaptive filtering and masking. I champion validation, observability, and production readiness for document intelligence, time-series, and graph-based AI initiatives.

I led end-to-end delivery of the company’s first AI-driven product, translating ambiguous business requirements into production-ready ML systems and delivering a public demo within three months.

I designed a modular evaluation framework for the full RAG stack and used benchmark results to guide model and architecture decisions. I evolved an early prototype into an agentic RAG system with dynamic tool selection and self-reflection, focused on scalability and continuous improvement.

I deployed and scaled the service for concurrent use with Docker, Azure App Service, and GitHub Actions, sustaining 99% uptime. I also contributed to a shared AI library and interactive endpoints, and piloted Microsoft Copilot MCP agents integrated with Teams.

I built and maintained large-scale data pipelines and more than 30 automated Databricks workflows to improve operational efficiency across teams.

I developed predictive maintenance and time-series anomaly detection systems with automated MLflow retraining, contributing to cost savings and new client contracts. I applied LLMs and NLP to service-report classification, safety issue detection, and retrieval-augmented information access.

I designed hybrid ML systems combining expert rules, time-series analytics, and NLP to improve alarm detection accuracy and response times. I also delivered customer dashboards and real-time monitoring solutions using Grafana, Azure Data Explorer, and Azure Data Factory.

I developed deep learning-based anomaly detection models for marine engine monitoring.

I contributed to digital and machine learning initiatives for maritime operational data and equipment performance.

I built trajectory prediction and collision avoidance models using AIS data and LSTM-based sequence modeling.

I developed visualization tools to support maritime navigation and safety analysis.

I built an NLP-based document classification system with OCR support for multilingual legal contracts.

I applied AI and language-processing methods to improve classification of document content.

GPA: 4.82 / 5. Graduated with distinction.

GPA: 3.83 / 4. Department Honors.