MealMatch AI

MealMatch AI is an early-stage startup building an AI-powered platform that transforms grocery shopping into a personalized, health-focused experience. As a Software Engineer on a team of 4+, I work across the React Native client and the Flask backend, owning cross-stack initiatives that span SDK integrations, authentication hardening, and security remediation alongside core mobile features like meal plan generation and budget tracking. The platform leverages LangChain and OpenAI to generate customized meal plans from 15,000+ recipes, factoring in dietary restrictions, health conditions, and budget constraints.

Highlights

• Owned end-to-end Customer.io SDK integration in React Native, taking a broken PR through diagnosis, refactor, and production launch: root-caused the SDK misuse blocking iOS device registration, consolidated 11 direct call sites into a single TypeScript service abstraction, instrumented 12+ lifecycle events, and wired identity cleanup into logout and account-deletion flows, covered by 37 Jest unit tests
• Hardened React Native and Redux authentication in response to recurring JWT signature errors: launch-time token expiration validation with automatic refresh, auth guards across 4 protected screens to prevent stale REST calls during logout, and a global response interceptor that catches 422 signature failures and surfaces a recoverable “session expired” flow
• Tightened Flask authentication middleware with defense-in-depth patterns: added user-existence validation to the token blocklist and refresh endpoints so deleted accounts cannot reuse outstanding sessions, and built a single-query SQLAlchemy blacklist-sentinel scheme that atomically revokes every session of a deleted user
• Identified and remediated 37 tracked security vulnerabilities across the Flask backend and React Native client: replaced hardcoded API keys and database credentials with environment variables, applied CORS origin restrictions across all REST endpoints, migrated from the Werkzeug development server to gevent for production, scoped cleartext HTTP traffic to debug builds only, and removed secret-containing files from git tracking
• Created an AI context and documentation system (structured YAML + human-readable guides) that reduced token usage by 40% across multiple agent runs and improved code change accuracy for both contributors and AI tools
• Built the budget-aware meal plan generation flow with proportional budget allocation across weekly, bi-weekly, and monthly planning periods, including the settings screen, overview display, daily cost aggregation, and Redux state management
• Developed the React Native meal planner UI including day/week navigation, circular progress indicators, and loading states aligned with Figma designs
• Built the user onboarding flow collecting health info, dietary preferences, and physical profiles, plus ZIP code geocoding for regional grocery pricing
• Built the feature usage limiting system with access gating to support subscription tiers
• Debugged cross-stack React Native and Flask integration issues, coordinating handoffs and ownership to cut feature turnaround from multi-week cycles to days

Technologies

Links