Raghavendranath Kandula is a software engineering leader and researcher with more than a decade of experience building and operating data platforms. His work focuses on financial document extraction, applied AI and production reliability. His research examines consistency checks for LLM-based extraction and silent failures, where a pipeline publishes an incorrect value without raising an alarm.
At Bloomberg L.P., he leads five engineers responsible for an earnings estimates extraction pipeline covering unstructured sell-side research and structured broker feeds, from acquisition through publication on the Bloomberg Terminal. Since joining Bloomberg in 2019, he has progressed through three role levels to Team Leader.
He initiated and leads the automated extraction of analyst ratings and price targets using large language models, the first production use of LLMs within his engineering organization. His responsibilities include technical architecture, post-processing validation, engineering safeguards for model promotion and drift detection, and phased rollout.
He also leads the migration of approximately 120 broker feeds, proposing shared publishing and directing a separate configuration-based redesign to simplify onboarding. His observability initiative supports a pipeline processing approximately 30,000 messages daily, with shared capabilities, including a circuit breaker, structured logging, alerting and synthetic health checks, adopted across all components owned by the team. He personally designed and built an analytics application that uses human corrections to identify extraction failure patterns. Earlier contributions include human-review selection criteria, legacy migration planning and onboarding automation that increased throughput from approximately five to 1,200 companies daily.
Previously, he built cloud communications components at Vail Systems and responded to production outages at Oracle. As a graduate research assistant in the University of Iowa’s Haim Laboratory, he applied computational and statistical methods to biological sequence data and co-authored a 2020 publication in mBio.
He holds a Master of Computer Science from the University of Iowa and a Bachelor of Technology in Computer Science and Engineering from Jawaharlal Nehru Technological University, Hyderabad. He also completed a postgraduate program in Artificial Intelligence and Machine Learning through the McCombs School of Business at The University of Texas at Austin. He has taught business-unit courses at Bloomberg and mentors engineers.