Sunil Oṁ Sharma

About Sunil Oṁ Sharma

I am a Senior AI Systems Architect with over 18 years of experience designing and delivering mission-critical enterprise systems. My focus today is architecting production-grade Generative AI platforms that operate securely, reliably, and at scale within complex enterprise ecosystems.My work sits at the intersection of AI architecture, distributed systems, and business execution. I design end-to-end AI systems spanning data foundations, semantic retrieval, reasoning orchestration, structured evaluation, observability, and production reliability ensuring foundation models translate into measurable enterprise impact.I specialize in building applied Generative AI systems that move beyond experimentation into durable production environments. This includes RAG pipelines, multi-agent orchestration workflows, LLMOps platforms, and AI-enabled engineering systems integrated with enterprise data platforms and event-driven architectures.I bring advanced hands-on expertise with Claude Code, using it as a high-leverage AI engineering platform to design and build production-grade capabilities across observability, reusable RAG patterns, multi-agent systems, and modernization of data pipelines. I have used Claude Code to accelerate repository-aware development, architecture generation, implementation planning, refactoring, debugging, testing, documentation, and workflow automation. My work includes reusable engineering patterns for RAG, orchestrated agent workflows, evaluation and guardrail layers, production observability, and scalable data/AI pipeline modernization.My expertise lies in treating AI as a governed subsystem within distributed architectures, not as a standalone feature. I focus on designing systems that are resilient, observable, economically sustainable, and aligned with compliance and security standards.I have also architected and delivered cloud-native, high-availability systems for mission-critical financial and enterprise platforms, including distributed transaction flows, real-time data pipelines, and resilient integration architectures. My work spans continuous delivery, observability, fault tolerance, performance, security, and developer experience.I have also contributed to enterprise GenAI adoption through internal AI education and enablement efforts, helping teams understand, adopt, and apply Generative AI in practical engineering and business settings.I am driven by building AI systems that are technically rigorous, ethically grounded, and capable of delivering transformative impact at enterprise scale.

Experience

1. Sr Ai Systems Architect Senior Lead Ai Software Engineer

   Capital One

   Jan 2024 — Present · Chicago, IL, US

   Architected and led development of an enterprise GenAI platform that unified fragmented data and documents into a secure, scalable intelligence layer using LLM orchestration, vector search, and API-driven retrieval. Designed fault-tolerant agentic architectures integrating enterprise systems with LangChain/LangGraph, advanced RAG patterns, and guardrails including PII redaction, hallucination evaluation, and model lineage, enabling secure natural-language access to enterprise insights.Drove platform optimization, model evaluation, and multi-agent workflows to reduce inference latency and cost while enabling autonomous reasoning, reporting automation, and context-aware insights at scale.Identified gaps in manual interpretation of regulatory and security requirements in FRD/PRD documents and proposed an AI-driven compliance gateway to automate enforcement before production deployment. Designed and implemented a multi-agent workflow to analyze requirements, validate compliance, and generate policy-aligned code integrated with enterprise SDKs, with guardrails, evaluation checks, and governance controls for auditable compliance.Led architecture and delivery of a natural language analytics platform with a specialized Data Product Interface that unified multiple enterprise systems and simplified access to financial data, accelerating analytics workflows, improving data accuracy.Architected and deployed resilient event-driven data pipelines on AWS Serverless with auto-failover, monitoring, and automated recovery, supporting millions of credit card transactions daily. Designed an Event-Driven Data Mesh Architecture for Card Core and built the Card Instruments SDK integrating card processing with accounting, risk, and compliance platforms, while championing serverless, streaming, and event-driven patterns for scalability, resilience, and observability.

Education

• MIT Professional Education

  Applied Data Science, Artificial Intelligence

  2025

• MIT Professional Education

  Artificial Intelligence, Advanced Reinforcement Learning

• Periyar University

  Bachelor of Computer Technology

• Anna University

  Master of Computer Applications

• Carnegie Mellon University

  Master's in Generative AI & Large Language Models (In Progress)