I’ve been building and scaling AI/ML solutions on Adobe Experience Platform, designing real-time identity resolution and feature pipelines to power churn prediction and personalization models. I’ve worked on LLM-driven content generation and intelligent document processing, making workflows more efficient and reliable. I’ve also built RAG pipelines, integrated foundation models like GPT and Adobe Firefly, and set up end-to-end monitoring and retraining frameworks in Azure to ensure everything runs smoothly. Overall, I focus on making enterprise AI workflows smarter and easier to use.

I built an LLM-powered system to extract key information from existing documents and automatically populate digital forms with high accuracy, using context-aware prompt strategies to handle diverse form structures and inputs. I developed a robust pipeline for processing unstructured and semi-structured data, leveraging Python libraries such as PyPDF2, PDFPlumber, and PyMuPDF, and implemented validation and confidence-scoring mechanisms to ensure data quality and reliability. Additionally, I developed an immersive Unity 3D VR simulation to visualize emergency exit routes in real time, integrating realistic physics, lighting, navigation meshes, and custom C# scripts to dynamically guide users. I conducted usability testing to evaluate navigation clarity and training effectiveness, enhancing the overall safety training experience.

I built and scaled ML pipelines on Google Cloud, processing large volumes of patient EHR data to deliver faster insights for clinicians. I developed and fine-tuned time-series, ensemble, and deep learning models for early detection of conditions like sepsis, integrating real-time scoring with patient vitals and imaging data to support clinical decision-making. I also engineered feature pipelines, containerized ML services, and monitoring frameworks to ensure models ran reliably, remained explainable, and complied with healthcare regulations. Working closely with clinicians and IT teams, I focused on turning complex data into actionable insights that improved patient care.

I designed and deployed real-time fraud detection models on high-volume credit card transactions, improving accuracy while reducing false positives and enhancing the customer experience. I built anomaly detection pipelines, engineered advanced behavioral and transactional features, and developed scalable end-to-end ML workflows to process large datasets efficiently. I collaborated closely with fraud and risk teams to integrate ML scoring into real-time systems and built customer-facing solutions like intelligent alerts and chatbots. I also set up monitoring, explainability, and automated CI/CD pipelines to ensure models ran reliably, stayed compliant, and adapted proactively to emerging risks.