Sayan Banerjee

Research Analyst

Hevo Data•  June 2024 - December 2024•  Bangalore

At Hevo Data, I managed and optimized 80+ ETL/ELT pipelines, ensuring seamless data ingestion with 95% error reduction through rigorous testing and process improvements. I collaborated with engineering teams to develop and enhance data connectors using REST APIs and secure TCP/IP protocols. Additionally, I led the design and implementation of a Snowflake Pricing Calculator, driving a 25% increase in MQLs and improving user engagement. I also developed a context-aware Support ChatBot, deployed across 50+ Hevo blogs, boosting customer retention by 15%. My role involved real-time data monitoring, troubleshooting pipeline issues, and implementing scalable solutions to enhance system reliability and efficiency.

Research and Development Intern

Samsung•  November 2023 - May 2024•  Remote

At Samsung, I developed an anomaly detection system using RoBERTa-Large, achieving 86.3% accuracy in identifying app policy violations and suspicious user behaviors. I implemented a RAG-based architecture leveraging LLMs like Llama 3.2 1B to build scalable AI-driven solutions. Additionally, I created automated monitoring systems using Python and SQL to analyze 10,000+ user interactions, reducing false positives by 40% and improving incident response efficiency. I also designed a data generation pipeline using Mistral 3B and Grok, enabling synthetic dataset creation for robust application testing and validation. My role focused on applying cutting-edge AI techniques to enhance security, compliance, and business decision-making.