At ArcelorMittal Nippon Steel India, I design, build, and operate production ETL and streaming pipelines for industrial time-series data. My work improved plant reporting efficiency by 20%.
I developed a real-time ingestion and alerting service handling 10,000+ sensor data points per second. I also implemented data validation checks and optimized SQL queries, improving retrieval speed and accuracy by 25%.
At Ebrandz LLP, I developed Python extraction pipelines and data quality scripts for web properties. I integrated AI/NLP-assisted analysis into production reporting, improving diagnostic accuracy by 15%.
I built an LLM/NLP-based SEO analysis system that detects ranking drops, indexing issues, and anomalies, with an interface for non-technical users to query insights. I also built a LightGBM drug side effects classification model that reached 90.4% accuracy, and published the findings in the International Journal for Multidisciplinary Research.
