In this role, I design and deploy production-grade AI systems centered around large language

models, agentic workflows, and tool-using agents. My responsibilities include architecting end to-end GenAI solutions, building retrieval-augmented generation pipelines over enterprise data,

and integrating LLMs with internal APIs, databases, and automation services. I work extensively

with Python, containerized services, Kubernetes, and cloud infrastructure to ensure these AI

systems are secure, scalable, and maintainable. A key part of my role involves translating

ambiguous business needs into deployable AI solutions while collaborating closely with

product and engineering teams.

I have delivered LLM-powered systems that reduced manual analysis and decision workflows

by over 40% across internal teams. I improved response accuracy and reliability by

implementing RAG pipelines and prompt evaluation mechanisms that materially reduced

hallucinations in real usage. I productionized AI services as low-latency APIs with enforced cost

controls and access isolation, and introduced monitoring loops for prompt, agent behavior, and

failure analysis. My work established reusable internal patterns for safely deploying agentic AI

systems at scale.

At Deloitte, I worked on building and productionizing machine learning–driven data pipelines

that supported both analytical and operational use cases. My role focused on designing

feature-ready data models, building scalable Python and SQL pipelines, and ensuring data

reliability for downstream ML workflows. I collaborated closely with data scientists to

operationalize experimental models into stable production assets, ensuring they could be

consumed reliably by business stakeholders.

I enabled faster model iteration cycles by designing clean, feature-ready transformations and

integrating ML outputs into dashboards and downstream systems. I improved reliability by

implementing data validation and monitoring within pipelines, reducing downstream reporting

issues. My work helped teams transition from experimental ML efforts to production-ready,

enterprise-grade solutions that delivered consistent insights within existing workflows

In my early career at Accenture, I built and automated backend workflows using Python and

SQL, supporting data pipelines and recurring client deliverables. I worked on developing and

maintaining production systems that handled multi-source inputs, ensuring consistent data

availability for analytics and reporting use cases. This role strengthened my foundation in

software engineering, modular design, and production accountability.

I reduced manual processing through automation, improved pipeline reliability, and delivered

cloud-deployed solutions under real client timelines. I built reusable utilities, standardized

ingestion logic, and collaborated across teams to ship systems used by real clients. This

experience laid the groundwork for my transition into AI and GenAI system ownership roles.