Eshan is a Core AI Software Engineer at Keystone, based in the Seattle office. He specializes in developing and scaling advanced software solutions that drive measurable impact for enterprise clients. In this role, he leverages deep expertise in Engineering, Generative AI, Machine Learning, and Cloud-native architectures to build intelligent products.
With a strong background in research and applied technology, Eshan previously engineered Moderna’s first enterprise-wide AI agent service, transforming internal processes across finance, clinical research, and supply chain management. At Moderna, he pioneered solutions using Retrieval-Augmented Generation, real-time NLP-to-SQL translation agents, and Kubernetes-based vector store pipelines on AWS infrastructure, collectively achieving operational accuracy exceeding 97%.
Prior to Moderna, Eshan was a Research Scientist at Northeastern University, where his groundbreaking work on transformer-based video segmentation set new benchmarks. His research advanced the field’s understanding of how attention mechanisms can mimic human perception. Additionally, as a Research Assistant at Khoury College of Computer Sciences, he co-authored a peer-reviewed publication on multimodal reinforcement learning frameworks for robotics, integrating visual and force data streams to enhance robotic precision and adaptability.
At Google Summer of Code, he contributed to Activeloop’s DeepLake platform by automating ingestion workflows for large-scale machine learning datasets, enabling users to ingest data with just a single line of code instead of hundreds, significantly streamlining the onboarding and scaling process for data practitioners.
Earlier in his career, Eshan developed enterprise workflow automation at Deloitte, leveraging advanced scripting on ServiceNow platforms to improve operational efficiency for U.S.-based clients. He also built scalable data analytics pipelines at PriceEasy AI, enabling real-time insights from complex geographic and economic datasets.
Outside of work, Eshan is an active contributor to the Generative AI community, collaborating on open-source projects and publishing academic research, driven by his passion for advancing AI’s practical applications in industry.