About
Me.

• Architected an enterprise multi-agent system with 5+ specialized autonomous agents to resolve complex L2 production incidents across distributed enterprise systems and data silos. • Slashed incident resolution time from 7–8 hours to minutes, drastically reducing operational downtime by automating investigation, root-cause synthesis, and remediation workflows. • Built an enterprise intelligent assistant that automates 100+ standard operating procedures (SOPs), providing verified citations and calibrated confidence scores to ensure reliable decision-making. (More...)

• Built an enterprise Multi-Agent NL2SQL system achieving 90%+ accuracy across 10,000+ queries with validation controls and response quality guardrails. • Designed LLM evaluation workflows with GPT and LLaMA models using RAG to extract structured insights from textual data. • Delivered demand forecasting for 5,000+ SKUs across Europe, generating forecasts across 7 levels of product hierarchy and multiple granularities, with 100,000+ models backtested monthly; improved WMAPE from 70% to 40%, reduced compute usage by 80%, and saved €10,000+ in Databricks costs. (More...)

• Engineered NLP analytics solutions and scalable ML pipelines on Azure Databricks for large-scale enterprise text processing. • Built automated topic discovery and Zero-Shot text classification (BART-Large-MNLI & LDA), eliminating days of manual data annotation. • Established robust MLOps practices with MLflow for experiment tracking, model versioning, automated selection, and reproducible evaluation. (More...)

• Built personalized product recommendation engines for curbside pickup marketplace platforms, improving customer experience through AI recommendations. • Designed semantic matching algorithms using advanced NLP to improve product search relevance and price intelligence. • Implemented AutoML pipelines (Regression, Classification, Clustering) for an enterprise No-Code AI platform, delivering production FastAPI microservices and Dockerized AI models. (More...)