Excited to lead the development of Amex's knowledge management tool for hotel bookings using cutting-edge technology like GPT-4 and RAG. Collaborated with stakeholders to integrate with Microsoft's Co-Pilot API, resulting in improved agent efficiency and customer satisfaction.Key Contributions:Building Amex’s first knowledge management tool for hotel bookings via concierge using Retrieval Augmented Generation (RAG)● Working with agents, operations, and model governance stakeholders to drive continuous enhancements of the chatbot● Integrating our internal knowledge management bot with Microsoft’s Co-Pilot API and to build a single travel Q&A solution● Increased call-to-bookings ratio in our pilot population (1.2 x higher than average) by serving up relevant hotel informationto agents who are helping customers book their travel plans● Double agent satisfaction with the tool from an NPS of 19% to 45% by prioritizing enhancements based on their needs,without compromising business or operational goals and savings projections● Increased end-customer (concierge callers) NPS by 5% points as a result of increased agent efficiency due from the tool● Oversaw user management tasks, including developing trainings to onboard new users and expanding the scope of the pilot

In my role as a Technical Product Manager at American Express, I spearheaded the development of groundbreaking products utilizing Large Language Models for customer service use-cases. This included implementing ML-based solutions to enhance in-app mobile search functionality, resulting in substantial cost savings. Additionally, I led the creation of an RPA tool with computer vision for document verification, saving $500,000 during the pilot stage.Key Contributions:● Pioneered the development of Amex’s first products (POCs & MVPs) which leverage Large Language Models (GPT-4, and LLaMA 2)● Developed the first internal intent classification & speech-to-text models for phone servicing, processing 500k requests amonth, saving $250,000 annually in vendor costs● Designed and implemented an ML-based solution to power Amex’s in-app mobile search to a pilot population of 800k users● Used question and answering techniques via search to deflect 8,000 calls a month and automate 50% of servicinginteractions that originate from the app, resulting in a monthly savings of $12,000● Enabled the company’s first connection to a SaaS LLM (GPT 4) for use in developing product proofs-of-concept● Built an LLM and vector-based POC for suggesting responses and smart-compose sentence completions to servicing agents● Oversaw the development of an RPA tool with computer vision for document verification, saving $500,000 during pilot stage● Devised the enterprise-wide strategy for prompt versioning and management as part of the model governance pipeline

In my role as a Product Owner at American Express, I devised a platform strategy for MLOps, reducing model deployment time by 50%. I piloted automated servicing for email inquiries, saving $80,000 in personnel costs. Additionally, I developed multilingual chatbot automation models, automating over 40k conversations per month. Key Contributions:● Devised a platform strategy around comprehensive MLOps, reducing model deployment time from 2 weeks to 1 day● Piloted automated servicing capabilities for email inquiries, saving $80,000 in personnel costs● Developed the first in-house French, Spanish, and Japanese chatbot automation models, automating >40k convos/month● Hired a product analyst responsible for implementing Amex’s first HITL framework, used to label >360k data points per year● Added support for bias reduction during labeling by automating the evaluation of inter-annotator agreement

In my role as a Product Analyst at American Express, I oversaw the development and integration of an ML-based sentiment analysis layer for real-time conversation monitoring. I collaborated with Engineers & QA to implement DART testing & zero-downtime (SRE) across the ML platform. Additionally, I led the analytics and data mining operation for Amex’s customer service COVID Response Initiative, ensuring efficient and effective customer support during challenging times. Key Contributions:● Oversaw the development and integration of an ML-based sentiment analysis layer for real-time conversation monitoring● Worked with Engineers & QA to implement DART testing & zero-downtime (SRE) across the ML platform● Transitioned the servicing platform from RNN (spaCy) to Transformer (BERT) models, increasing accuracy and automation● Led the analytics and data mining operation for Amex’s customer service COVID Response Initiative● Conducted market research on chatbots in financial services to inform the product roadmap● Collaborated with the data science team to train Amex’s first in-house NLP model for a servicing chatbot

• Built, and currently manage, all of the firm’s digital service operations (all user traffic, readership ana-lytics, client’s user data, publishing, and weekly, automated processes)• Creation and analysis of user behavior, web traffic, and campaign data into reports using Tableau• Set up and maintained user, web, and campaign data pipelines into SQL and Big Query databases• Automated and integrated all payment software for accounts receivable and payable• Integrated data from Google Analytics, Squarespace, and all major content distribution platforms

• Conducted market research on chatbots in financial services to inform the roadmap for the product • Creation of automation reports using Tableau and Splunk for use in AB testing and decision making • Collaborated with the data science team to train a new NLP model for the chatbot• Conceived and oversaw the process for integrating a new sentiment analysis layer in our ML models • Participated in the automation and implementation of customer journeys for both web and chatbot

I worked as a growth hacker at Seashells. My aim was to increase our social media following, optimize our search engine rankings (SEO), and create content marketing. Using Crowdfire, I used hashtags, Twitter keywords, and Facebook Analytics to find users who would love to use the app for credit card saving. In the same vein, I use Google Analytics and Adwords to increase our Google rankings. Finally, I did a little business development by engaging with journalists and publications who were interested in spreading the news about Seashells' savings.

In my role as Lead Instructor at Mathnasium, I spearheaded the development of a comprehensive SAT curriculum, integrating quick practice tests and in-class sessions to enhance student performance and engagement. By leveraging my tutoring experience and research skills, I successfully tailored the curriculum to meet the evolving needs of students preparing for the SAT exam in Tampa, Florida.