Agentic Conversational AI & NLU Strategy Leader

Amanda Mahnaz Yazdani

Port Jefferson, NY 443-***-**** ******.******.*******@*****.*** linkedin.com/in/amanda-yazdani

Conversational AI Leader Agentic AI and LLM Strategy NLU/NLP Expertise Enterprise AI Implementation

PROFESSIONAL SUMMARY

Conversational AI leader and Ph.D. linguist with 5+ years of progressive experience spanning conversational AI design and engineering (CVS Health), natural language understanding (NLU) training and conversational design (Dexcom), and agentic AI solution ownership (Parloa). Combines NLU/NLP, large language model (LLM), and agentic AI depth with measurable business impact — NLU accuracy raised from 35% to 98%, business savings up 20%, call containment 8% above target — and builds the evaluation protocols and key performance indicator (KPI) frameworks that make conversational AI quality measurable. Leads client discovery, coordinates cross-functional CX, IT, engineering, vendor, and business stakeholders, and mentors junior developers, grounded in Brain & Cognitive Science and Linguistics.

KEY ACHIEVEMENTS

AI performance optimization: Improved NLU intent-classification accuracy from 35% to 98% for Dexcom's corporate chatbots, generating approximately $300K in annual business savings (CDI Services / Dexcom)

Cost reduction and containment: Increased business savings by 20% and held call containment 8% above target for CVS Health Aetna's conversational AI program (approximately 37,000 calls per month)

Business-impact measurement: Designed a KPI framework quantifying chatbot cost avoidance, demonstrating a roughly 50% reduction in live-agent call volume for Dexcom's Marketing and Customer Care department

Production delivery leadership: Led successful overnight go-live deployments of new features and hotfixes at CVS Health Aetna; supported go-lives at CVS Health Caremark and production deployments at Dexcom

Talent development: Onboarded, trained, and mentored four junior developer hires (CVS Health Aetna) and three new development hires (Dexcom); authored client-education, knowledge-transfer, and best-practice documentation

Innovation: Winner, 2023 CVS Generative AI Hackathon

Enterprise solution ownership: Lead agentic AI requirements discovery with enterprise clients and implement customer-service solutions on Parloa's AI platform across three enterprise-scale implementations, defining AI evaluation protocols

PROFESSIONAL EXPERIENCE

Lead AI Agent Architect

Parloa Inc. February 2026 – August 2026

Lead discovery of agentic AI solution requirements with enterprise clients and implement customer-service solutions on Parloa's enterprise conversational AI platform across three enterprise-scale implementations

Simplify complex workflows and business processes into reliable, user-aligned LLM prompts through prompt engineering, translating business requirements into dependable AI agent behavior

Develop structured simulation and evaluation protocols — an AI evaluation framework for agentic AI agents — setting repeatable quality standards for testing, validation, and quality assurance

Partner cross-functionally with CX and IT on complex integrations, defining API contracts and data types

Beta-test new Knowledge Base platform features with lighthouse clients

Author internal and external best-practice and how-to guides, codifying implementation standards

Conversational Designer

TEKsystems / Dexcom June 2025 – December 2025

Developed proactive content strategies for continuous improvement of Dexcom chatbot interactions

Designed a key performance indicator (KPI) framework that quantified chatbot cost avoidance and demonstrated a roughly 50% reduction in live-agent call volume for the Marketing and Customer Care department

Analyzed conversational data and performance metrics to identify trends and end-to-end user-experience improvements

Implemented and maintained NLU systems for customer-support chatbots; trained and fine-tuned NLU models for intent classification with machine-learning techniques and tooling

Collaborated cross-functionally with CX and IT teams to ensure seamless integration of chatbot enhancements

Conversational AI Trainer

CDI Services / Dexcom March 2024 – February 2025

Improved NLU intent-classification accuracy from 35% to 98% for Dexcom's corporate chatbots through chatbot training and optimization, delivering approximately $300,000 in annual business savings

Designed, developed, and improved NLU systems — intents, entities, and dialogue flows — for corporate chatbots; evaluated NLU performance to identify optimization opportunities

Used and adapted Python tooling to evaluate chatbot performance and quality

Partnered with engineering teams to ensure production deployments

Developed client-education and knowledge-transfer documentation; trained and mentored three new development hires

Conversational AI Engineer

CVS Health Aetna February 2022 – February 2024

Implemented AI enhancements for a conversational AI program handling approximately 37,000 calls per month (about 444,000 calls per year), increasing business savings by 20%

Increased and maintained call containment 8% above target; developed additional AI features reflecting business rules in conversational logic

Led successful overnight go-live deployments of new features and hotfixes

Managed regular communication with vendors, business stakeholders, and IT (technical and non-technical)

Onboarded, trained, and mentored four junior developer hires on process and procedure

Won the 2023 CVS Generative AI Hackathon as team advocate and presenter, refocusing the team on the core problem; the winning tool helps physicians get prior-authorization requests accepted more easily

Conversational AI Designer

CVS Health Caremark July 2021 – February 2022

Developed voice user interface (VUI) speech-to-text and text-to-speech (STT/TTS) systems in IBM Watson

Expanded the Python API natural language processing (NLP) tooling used by conversational engineers

Identified and fixed critical errors in dialogue flows and conversational responses

Supported successful overnight go-lives of new features and hotfixes; maintained code and source-control repositories and created documentation for development and CI/CD processes

EDUCATION

Ph.D., Linguistics — Stony Brook University, NY

Post-baccalaureate, Brain & Cognitive Science, Linguistics, and Dance — University of Rochester, NY

Bachelor of Arts (B.A.), cum laude, Japanese Language & Russian Language — University of Rochester, NY

Fine and Studio Arts, Portraiture — Atelier of Flowerfield / McEvoy Atelier, NY

SKILLS

Leadership and Strategy: AI Strategy, AI Assistant Development and Delivery, Conversational Design, Client Discovery, Cross-Functional Collaboration, Vendor and Stakeholder Communication, Agile Processes, Knowledge Transfer, Developer Training and Mentoring

Generative AI and Agentic AI: Generative AI, Large Language Models (LLMs), Prompt Engineering, Agentic AI and AI Agents, Simulation and Evaluation Protocols (AI Evaluation Frameworks), Parloa AI Platform

NLU and NLP: Intent Classification, Entity Recognition, NLU Model Training and Fine-Tuning, Chatbot Training, Chatbot Optimization, Dialogue Flow Design, Speech-to-Text (STT), Text-to-Speech (TTS), VUI Development

Chatbot Analytics and Measurement: KPI Framework Design, Cost-Avoidance Metrics, Conversational Analytics, Conversation-Quality Measurement, Qualitative and Quantitative Analysis, SQL, BigQuery, Splunk, Tableau

Enterprise Conversational AI Platforms and Tools: IBM Watson, Google DialogFlow, Voiceflow, Python, JavaScript, HTML/NodeJS, APIs, API Management, Database Management, Source Control, CI/CD Documentation

Languages and Other: English, Japanese, Russian, Mandarin, French; Microsoft Office, Voice Talent, 3D/2D Illustration

AWARDS & HONORS

Winner, 2023 CVS Generative AI Hackathon — CVS Health

Contact this candidate