Jainill sheth

• Owned end-to-end design and delivery of large-scale data pipelines supporting fraud investigation, financial
analytics, and AI-driven workflows in production environments.
• Built and operated real-time and batch ingestion pipelines using Kafka and Dataflow to stream data from
Guidewire systems into cloud storage and analytics platforms.
• Developed retrieval-augmented generation (RAG) systems using LLMs such as Gemini, OpenAI, and Claude
(Anthropic), enabling automated legal and financial document analysis with human-in-the-loop feedback.
• Implemented AI agent orchestration pipelines using LangChain and LangGraph to support multi-step, agent
based workflows.
• Applied prompt engineering and evaluation techniques to enhance model accuracy, reduce hallucinations,
and ensure safe and high-quality outputs for legal, financial, and insurance use cases.
• Containerized data and AI workloads using Docker and deployed scalable services on Kubernetes across
cloud environments.
• Designed scalable ETL workflows in Python, SQL, and PySpark using Airflow for orchestration, ensuring
reliability, monitoring, and recoverability in production systems.
• Optimized distributed Spark workloads and SQL queries using window functions and performance tuning,
improving query performance by ~25%.
• Collaborated closely with business stakeholders, analysts, and compliance teams to understand end-user
workflows and ensure data solutions met real business needs.
Tools used: GCP, Airflow, Databricks, Python, PySpark, SQL, LLM, Kafka, LangChain, LangGraph, GitHub

• Led migration of enterprise ETL pipelines from Informatica and on-prem Hadoop environments to cloud
native architectures on GCP using BigQuery and Terraform.
• Designed and maintained scalable batch data pipelines using Spark, Airflow, and Databricks, supporting
supply chain and financial analytics.
• Built CI/CD pipelines using Jenkins to automate deployment and reduce manual operational effort.
• Refactored Hive-based transformations into optimized BigQuery SQL workflows, significantly reducing
pipeline runtimes.
• Worked across distributed systems and heterogeneous data sources, including Oracle, SQL Server, and third
party operational systems.
• Partnered with data architects, business analysts, and migration teams to drive data discovery, clarify
requirements, and deliver reliable production solutions.
Tools used: GCP, Spark, Airflow, Databricks, Hive, Oracle, Cloudera, Hadoop, Jenkins (CI/CD), GitHub