• Developed and launched Knoumi, a full-stack mental health platform providing AI-powered therapy sessions and intelligent support group matching

• Built and deployed 3 AI therapist personas using OpenAI’s GPT-4 Turbo Chat Completions API, leveraging system prompts and Firestore for personalized, context-aware AI therapy sessions

• Designed and implemented an intelligent support group recommendation system, matching users with tailored support group suggestions based on AI-driven analysis of therapy conversations

• Currently in beta testing with users who also engage in traditional therapy, gathering insights to refine AI responses and enhance user experience

• Spearheading product strategy, branding, and future development to bridge the gap between AI and mental health support, with a vision for expansion in Japan and global markets

• Conducting API and load testing using K6 and Firebase Emulator to ensure high performance and stability, achieving a 99% query success rate and 0% authentication failure rate

Current Tech Stack: React, Typescript, Node.js, Nest.js, Firestore, Firebase, OpenAI API

Team: Amazon Ads Org - Shopper Panel Team

• Worked in Machine Learning centric team on Amazon Shopper Panel to perform receipt image validation, extraction of data from receipts, and mapping data to consumer products in order to facilitate Amazon advertising efforts

• Implemented pre + post processing functions in AWS Lambda for custom SageMaker Ground Truth jobs to serve OCR text and/or receipt images, as well as consolidate and process results to be stored and used by data science teams

• Developed front end templates and deployed custom Ground Truth annotation jobs utilizing data from the Shopper Panel app to support various ML / data science team requirements for retailer mapping and product pricing initiatives

• Converted various SageMaker Notebooks for streaming labeling job creation into Lambda functions, optimizing workflows and significantly reducing AWS costs by improving resource efficiency

• Led frontend development of new greenfield B2B web application by creating full project architecture, managing team of engineers across international timezones, optimizing code continuously, and ultimately deploying for general availability

• Modernized multiple existing flagship app microservices with Angular to React, Redux, TypeScript, and Material UI

• Revamped pre-acquisition features and microservices to align with Dun & Bradstreet central theme and new overarching application architecture, integrating prior standalone applications into new suite of products while eliminating tech debt

• Implemented new features in the Rev.Up ABX product such as ad campaign management, creatives uploading, budget forecasting and optimization, and target audience selection, allowing customers to effectively customize ad campaigns

• Worked as on-call engineer to diagnose pipeline issues stemming from integration, unit, and Selenium tests

• Developed multiple user and administrator interfaces with modular designs by utilizing React, Redux, TypeScript, and Material-UI for the ‘HELPO’ COVID-19 PCR test scheduling and hospital management web app

• Collaborated with back end engineering team to develop and test new APIs and resolve issues pertaining to legacy APIs

• Created new retailer-side features with Flutter/Dart and Firebase for the ‘HELPICK’ pharmacy e-commerce mobile app