Uri Yael

Built a cash-flow assistant that helps users track recurring payment plans, projected transactions, and settled financial movements through weekly summaries and AI-assisted entry creation. I worked on the back-end and product workflows behind cash-flow extraction, feature flags, subscriptions, user identity, and notification flows. The system included React-based user flows, shared front-end packages, PostgreSQL-backed services, and Signal Harbor, an interaction layer for turning user messages into structured actions.

Built Olaid, a health application that lets users create and manage an emergency medical profile that can be shared quickly when needed. The app focuses on practical emergency access to relevant health information, including patient details, medical data, weight tracking, and a public shared profile view.

A multi-tenant workforce and operations platform built as a Python back end paired with a TypeScript web portal. I worked full-stack across the codebase, delivering feature modules end-to-end: An AI-powered WhatsApp agent (built on the OpenAI Agents SDK) that handles shift confirmations and contact lookups, shift-candidate proposal and matching flows, contact groups with broadcast messaging, a skills/competency system, a knowledge graph linking people, places, and risks through typed relationships, governed social-post voting workflows, and HR workforce-hours dashboards. I designed the strict organization-isolation model that keeps each tenant's data separate, implemented soft deletes and explicit transactional boundaries, and drove a super-admin feature flag and permissions system to address different organizational needs and support staged rollouts.

Built Signal Harbor, a communication gateway that connects WhatsApp conversations with integrated applications such as Olaid and Velora. The system receives inbound WhatsApp text, images, files, and voice messages, preserves conversation context, uses AI to determine the intended action, and routes the request to the correct back-end service.

Signal Harbor supports practical workflows such as validating a user’s phone number via WhatsApp, extracting structured data from audio or text, relaying AI-assisted inbound messages to Olaid and Velora, and safely handling files or media across service boundaries. I also implemented deployment and operational safeguards using GitHub Actions, systemd services, environment management, and persistent WhatsApp session storage.

Built and operated a mental health platform that helps people in Paraguay find psychologists and healthcare professionals through structured provider profiles. The platform supports profile creation and update workflows, including professional biographies, specialties, care modalities, locations, contact information, and approval flows. I worked across the product, back-end, deployment, and operational side of the platform, turning a fragmented discovery problem into a searchable, maintainable directory for mental health services.