Ganesh Adarkar

At ZeroOne AI, I architected IDEA, an end-to-end document processing platform, and built scalable extraction pipelines processing 100K+ documents daily. My deep learning models for engineering-drawing layout segmentation achieved 95% prediction accuracy.

I also built a LangGraph multi-agent system for an edtech platform with 100K+ learners, autonomously resolving 65% of tier-1 tickets within 90 days. I developed a RAG pipeline over help-centre content and historical resolutions that achieved 88% retrieval precision across 700 labeled production tickets.

Previously at Surreal AI, I owned real-time ETL pipelines for a 500K+ user chatbot, supporting decisions that drove a 30% uplift in engagement. At DeepCloud Solutions, I delivered Power BI and SQL reporting that reduced supply-chain stockouts by 18%, while XGBoost-based anomaly detection reduced data inconsistencies by 20%.

I work across Python, SQL, cloud infrastructure, MLOps, machine learning, NLP, and generative AI, partnering with product, engineering, analytics, and commercial teams to turn business questions into operational systems.