Free D.

Architected a self-healing scraping system on GCP Cloud Run combining AI Agents and cascading fallback logic to maintain 100+ real estate pipelines with zero manual intervention. Built multi-agent orchestration using MCP (Model Context Protocol) to coordinate AI agents for extraction, classification and data quality validation.

Designed advanced RAG pipelines (Hybrid Search + Re-ranking + Vector DB) for context-aware LLM responses over domain data. Fine-tuned video and language models via RunPod & Hugging Face; applied Model Editing for dynamic LLM updates without full retraining. Deployed microservices on GCP Cloud Run with full CI/CD via GitHub Actions and Docker.

Designed and maintained secure ETL pipelines on Skywise for critical military aerospace data (sensors, maintenance, operations). Built Spotfire KPI dashboards for real-time predictive maintenance monitoring across NH90 helicopter fleets. Automated workflows via Python/VBA scripts; managed CI/CD orchestration for pipeline reliability. Fine-tuned an open-source transformer model for aerospace technical document translation.

Architected a datalake integrating STT, OCR and NLP to process audio and textual coaching interactions at scale.

Deployed MLOps (Docker, Kubernetes, AWS) with real-time Kafka data flows and Spark/Hive processing.