Ismail Vanak

Developed U-net TensorFlow model for satellite image segmentation, incorporating Meta's Segment Anything. Engineered AI agents using LLM for web data extraction. Collaborated on RAG for chatbots. Orchestrated image segmentation pipelines on Google Cloud with Docker. Driving innovation in AI, ML, and MLOps methodologies.

Specializing in ML, DL, and NLP, I've innovated in agreement analysis, training models with TensorFlow's Universal Sentence Encoder. Led projects in MLOps, scaling pipelines on Azure, and deploying models on Azure ML Studio. Improved Python pipelines for legal contract attribute extraction by 50%. Led Presales, delivering 150+ agreement metadata for implementations, and played a key role in code review, testing, and CI/CD for AI pipelines.

Python expert specializing in data ingestion and ML pipelines, proficient in SQL and ETL. Predicts turbine outage and revenue from SAP data stored in denormalized PostgreSQL Data Lake. Implements job log analytics using SQL and Talend, pioneers Talend-based Data Ingestion Framework for automation. Skilled in Tableau for Grid KPI dashboards, leads as Data Governance Lead, delivering 28 KPI automation projects. Creates ETL jobs and Excel reports for data quality checks.