Kiran Kumar Ramanna

Kiran Kumar Ramanna’s fourteen-year career sits at the practical frontier where artificial intelligence becomes enterprise infrastructure. Across roles spanning ServiceNow and Apple, his work has consistently treated AI not as a demo artifact but as a production discipline—measured in deflection rates, latency budgets, safety controls, and reliability metrics. His focus has been to build systems that users can trust at scale: conversational platforms that route work correctly, retrieval pipelines that cite sources, and agentic frameworks that operate inside real governance boundaries.

At ServiceNow, Ramanna currently serves as a Senior Staff AI Engineer (since March 2024), working on the Agentic AI framework that powers conversational orchestration across the platform. The problem domain is not simply “answer generation,” but controlled, auditable enterprise behavior: retrieval-augmented generation with safety guardrails such as PII redaction and profanity filtering, hybrid search pipelines that blend embeddings retrieval, reranking, and LLM-generated answers, and personalization informed by Knowledge Graph signals and user-profile context. His work includes building evaluation harnesses, implementing citation mechanisms, and defining fallback strategies—critical design choices when AI must remain useful under uncertainty rather than fail silently or hallucinate.

From November 2021 to March 2024, as a Staff Machine Learning Engineer, Ramanna advanced Conversational RAG implementations and co-developed training-data strategies aligned with ServiceNow’s CoreLLM effort responsible for NowLLM. In that period, he built the Advanced NLU Workbench to enable multi-model batch testing and telemetry-driven performance measurement—work that moves AI development from anecdote to instrumentation, allowing teams to understand model behavior, regressions, and tradeoffs across real traffic conditions.

His platform impact at ServiceNow extends beyond AI. Since September 2020, he contributed as a Staff Software Engineer and became a founding engineer on Vaccine Administration Management during the COVID-19 response. He designed scheduling workflows, eligibility rules, and capacity management systems under urgent public-health timelines, partnering with government and healthcare stakeholders to support deployments across more than 100 organizations, including NHS Scotland. In this context, “enterprise reliability” carried real-world consequence: systems had to support public operations with predictable uptime and controllable rulesets.

Earlier at ServiceNow, beginning October 2016 as a Senior Software Engineer, Ramanna was part of the foundational build-out of Customer Service Management. He helped establish omnichannel interaction models spanning chat, messaging, voice, and email; built Agent Chat and Agent Workspace; implemented Advanced Work Assignment for skills-based routing; and enabled Virtual Agent capabilities with NLU and live-agent handoff. This work helped define how large support organizations manage conversations, route work, and maintain continuity across channels—an enterprise workflow problem that is structurally similar to today’s agentic orchestration, but solved initially with robust interaction and routing primitives.

Before ServiceNow, Ramanna’s work at ENACT Systems (2016) focused on distributed-energy platforms. He built secure partner APIs and integration layers for the solar ecosystem, delivering multi-tenant services with Spring and MongoDB, analytics for pipeline performance, and operational reliability through monitoring, tuning, and load testing on AWS. Earlier at Apple, he contributed to platform modernization efforts that emphasized modularity and scale. Leading modernization for the Radar project, he helped break monoliths into API-based microservices with versioned contracts, implemented near-real-time Oracle-to-MongoDB synchronization using Kafka and Akka patterns, and built Elasticsearch-backed search capabilities with indexing pipelines. He also delivered the MFi Licensee Portal using Spring, Hibernate, and Oracle, orchestrating complex workflows for third-party accessory manufacturers—experience that reinforced disciplined contract design and compliance-aligned workflows in large ecosystems.

Ramanna’s technical contributions have been recognized through a pending patent on Dynamic Orchestration of Multi-Agent AI Workflows and a research paper on PrEAM (Prompt Optimization using Evaluations by Automated Micro-judges). At ServiceNow, recognition includes Net New Innovation Day Winner (2024), Best Team Player awards (2021, 2023, 2024), a Customer Service Management Innovation Award for an MFA proof-of-concept, and repeated hackathon participation (2017, 2018, 2020). He has complemented industry delivery with formal upskilling, including Stanford’s AI Leadership Series and UC Berkeley’s Professional Certificate in Machine Learning and Artificial Intelligence.

Across systems, a consistent theme stands out: building AI with guardrails, governance, and observability. Whether enabling public-health workflows in vaccine administration or deploying conversational and retrieval systems inside enterprise platforms, his work emphasizes auditability, safe defaults, and measurable performance—core requirements when intelligent systems must be trusted by organizations and the public.