Natalia Elezovic

Hey, I'm Natalia.

I currently work on Conversational AI at Bank of America, building on earlier work on Google Assistant at Accenture and my studies at the University of Chicago. My goal is to shape the next generation of AI, turning advanced intelligence into experiences and capabilities that people actually trust, adopt, and rely on at scale.

I bridge technical and business worlds, align teams around a shared vision, and drive AI initiatives from strategy through production.

My long-term vision is to help build AI that benefits humanity: grounded in strong governance, shaped by culture, and designed to help people make better decisions in an increasingly complex world.

This website brings together my work across the business and technical dimensions of AI, highlighting selected experiences at the intersection of technology, strategy, and execution.

I write about responsible AI, the future of leadership, and my experience working at the intersection of business and technology. In my free time, I explore and write about neuroscience and linguistics.

Career

Professional Experience

Currently focused on enterprise AI strategy, governance, and AI product delivery at one of the world’s largest financial institutions.

Bank of America
June 2024 to Present
Chicago, IL
Senior AI Engineer  ·  Conversational AI Team Lead

Erica for Employees  ·  Enterprise Conversational AI Platform  ·  Web & Voice

01
AI Strategy & Stakeholder Leadership
15+
Business Verticals
55+
unique user journeys
Technical Advisory Internal Collaboration Stakeholder-facing Scalable AI
  • Led enterprise adoption of the Erica AI platform across multiple lines of business, supervising adoption of advanced AI capabilities across 15+ business verticals and overseeing the end-to-end delivery of 55+ unique user journeys through governance, implementation, and production deployment.
  • Served as strategic AI advisor to executive stakeholders, translating complex NLP, conversational AI, and UX considerations into business decisions and implementation roadmaps.
  • Advised senior leadership on voice AI adoption, defining ASR performance standards (WER, MOS), success metrics, escalation strategies, and release readiness.
  • Acted as a bridge between Product, Engineering, Governance, and Senior Leadership teams to drive alignment and accelerate delivery.
02
Cross-Functional Leadership
40+
Team Members
Team Leadership Mentorship Hiring Strategy Jira/Agile Frameworks Building Scalable Frameworks Building AI Solutions (Weekly Releases)
  • Mentored and guided cross-functional teams of 40+ members (AI engineers, product managers, scrum masters, and business stakeholders) across design, development, testing, and production, guiding end-to-end agile delivery and platform implementation.
  • Coached 15+ AI Engineers through 1:1 mentoring, delegation, and ownership opportunities, helping build internal AI expertise across the organization.
  • Established best practices, technical delivery frameworks, and operational processes during platform implementation, utilized daily by 7+ different teams.
  • Influenced hiring and resource planning strategies by identifying capability gaps and recommending critical technical and organizational investments.
03
AI Governance & Risk Management
4
Enterprise AI Models
Model Risk Management Audit Readiness AI Governance Routines GenAI / RAG POC Responsible AI
  • Led governance and risk oversight for four enterprise AI models supporting web and voice experiences, driving compliance, audit preparedness, and executive governance routines for responsible AI deployment at scale.
  • Partnered with Model Risk Management leadership to define frameworks balancing innovation, regulatory compliance, and business objectives, navigating complex governance challenges and driving enterprise-aligned outcomes.
  • Owned governance and model risk management for a newly launched AI model serving 15 business verticals, ensuring alignment with enterprise standards and regulatory requirements.
  • Supported GenAI/LLM Enhanced Search proof-of-concept initiatives from a governance and risk perspective.
04
Model Performance & Excellence
35%
Fewer Escalations
5–7%
Avg. Accuracy Gain Across LOBs
Model Optimization Dashboard Analytics Gen AI POCs RAG Infrastructure Success Metric Definition Thrive in Ambiguity
  • Owned the optimization strategy for four enterprise conversational AI models, establishing governance, prioritization, and continuous improvement frameworks.
  • Increased model accuracy by an average of 5–7% across channels (Technology, HR, General Support, etc.) and contributed to a 35–40% annual reduction in escalation rates, improving user experience, operational efficiency, and business outcomes.
  • Directed prioritization of Jira backlogs and strategic initiatives to improve conversational performance; leveraged individual team strengths to maximize delivery outcomes.
  • Designed and launched an internal analytics platform (Streamlit) to measure conversational success and user journeys, recognized in the Bank of America Innovation Challenge.
Accenture
February 2022 – August 2023
Dublin, Ireland
Natural Language Processing Engineer

Google Assistant  ·  Multilingual Enterprise NLP

52%
Cumulative Model Performance Increase
Cross-Functional Collaboration Model Fine-Tuning Technical Documentation Mentorship
  • Supported the optimization and evaluation of AI-driven language systems powering Google Assistant products across multilingual enterprise environments
  • Led model performance improvement initiatives, contributing to a 52% increase in precision through structured workflow analysis, annotation refinement, and AI model optimization
  • Collaborated with engineering and product teams to strengthen NLP pipelines, improve automation workflows, and support scalable AI deployment initiatives
  • Conducted large-scale annotation, QA, and evaluation workflows to validate AI-generated outputs, identify inconsistencies, and improve operational quality standards
  • Developed technical guidelines and documentation adopted across the EMEA region to standardize AI evaluation workflows and improve execution consistency
  • Supported AI governance, quality assurance, and operational review processes by documenting findings, validating outputs, and contributing to model performance assessments
  • Delivered structured feedback and business-aligned recommendations to support AI optimization, workflow scalability, and enterprise user experience improvements

Writing

Articles & Perspectives

Field notes from the AI industry, published on Medium.

Coming Up Next

Ancient columns representing governance and structure
The Two Pillars of Stable AI, Explained

Governance makes AI safe; monitoring makes it beneficial. This piece breaks down the two disciplines every leader needs to get right, with practical frameworks and the questions worth asking before the next initiative goes live.

Published

Speaking

Featured Talks

Panels, talks, and conversations at the intersection of AI, data science, and human-centered technology.

July 2026 Coming Up First Podcast
Podcast Guest · "Chat Applied Data Science"
University of Chicago · Master of Science in Applied Data Science

ChatADS is the official podcast of UChicago's MS-ADS program, featuring candid conversations on how data science, machine learning, and AI are applied to solve real-world problems. Topics span industry applications, career trajectories, emerging tech trends, and ethical data choices.

February 17, 2026
Guest Speaker · Inside Tech: Career Journeys Across the Industry
KTP x Career Advancement · University of Chicago

Panelist in a moderated discussion for UChicago's Kappa Theta Pi winter workshop series. Joined 5–6 professionals across product, engineering, data science, and strategy to share career journeys and advice for students navigating tech recruiting, followed by student Q&A and open networking.

2025 – 2026
Guest Speaker · Master's in Applied Data Science
University of Chicago · Applied Data Science Graduate Program

Invited guest speaker for the ADSP program, sharing industry perspective on enterprise AI, NLP, and the path from engineering into AI product leadership.

March 29 – 30, 2025
Hackathon Judge · Uncommon Hacks 2025
Department of Computer Science · University of Chicago

One of a select group of judges chosen for exceptional professional experience in tech. Evaluated 13 project submissions across a 24-hour sprint with 150+ undergraduate and graduate participants from across the US. Provided technical feedback, asked probing questions, and contributed to final prize deliberations.

Beyond Work

The Personal Side

The interests, values, and experiences that shape how I think about technology, people, and the world.

Coming Soon

More of Natalia, soon

A closer look at the person behind the work: ideas, interests, and what matters to me beyond AI and product.