Lucas – Python, React, Typescript

USSSA is a large-scale sports platform that enables organizations, event hosts, and participants to manage registrations, payments, and financial operations across multiple regions.

The platform supports complex multi-party payment flows where participants pay for events, organizations collect fees, and different stakeholders receive payouts based on configurable business rules.

The system integrates with Stripe to handle onboarding, payment processing, fund management, and payouts, ensuring a seamless and reliable financial experience for all users.

• Defined the system architecture using a structured approach starting with C4 diagrams (system context and container levels) to align stakeholders on high-level design
• Modeled critical financial workflows using sequence diagrams and state lifecycle definitions to ensure consistency across payment operations
• Designed database schemas and API contracts based on validated domain flows, ensuring alignment between business requirements and technical implementation
• Established a clear implementation roadmap derived from architecture artifacts, enabling efficient and predictable delivery of complex payment features
• Led the evolution of the platform towards an event-driven architecture with asynchronous processing for scalability and reliability

A popular EdTech tool used by over 2 million students to overcome academic obstacles by generating study notes, summaries, flashcards, and quizzes from lecture recordings. It offers an AI-driven, personalized approach to learning and problem-solving, with capabilities to help with productivity and focus.

• Led a major backend refactor from synchronous to fully asynchronous architecture, improving request concurrency and overall system throughput;
• Reworked critical API endpoints using async patterns to reduce blocking I/O and optimize latency;
• Implemented internal KPI dashboards for monitoring learning engagement, performance metrics, and system health;
• Designed observability layer for tracking AI response time, error rates, and usage patterns;
• Improved backend scalability to better support growing user traffic;
• Implemented MongoDB-based async data pipelines for storing AI interaction logs and engagement metrics;
• Built aggregation queries and indexes to support KPI dashboards and performance monitoring;
• Leveraged Claude Code with GSD as a meta-prompting layer to accelerate backend refactors, enforce architectural consistency, and streamline async migration workflows;
• Managed schema evolution and indexing strategies to improve query latency at scale.

An AI-powered beauty companion that helps users organize short-form beauty videos and identify products used by creators — without digging through comments. It allows users to save Instagram and TikTok beauty videos, automatically extract referenced products, and build a structured, searchable product library to make informed purchase decisions.

• Designed and implemented scalable scraping pipelines for Amazon, Sephora, Ulta and others beauty products sources
• Built resilient crawlers with data normalization and anti-bot handling strategies
• Structured extracted product metadata for downstream AI-driven matching and recommendations
• Implemented asynchronous data ingestion workflows for high-volume processing
• Improved product identification accuracy by enriching scraped datasets
• Implemented high-throughput async ingestion pipelines storing normalized product metadata in MongoDB
• Optimized indexes and query patterns to support fast product matching and AI recommendation workflows
• Utilized Claude Code combined with get-shit-done to rapidly iterate on scraping pipelines, optimize ingestion workflows, and standardize data modeling across sources
• Structured semi-structured product data models for flexible ingestion from multiple sources

A multi-tenant marketplace platform connecting artists and venues, managing the full proposal-to-payment lifecycle.

Designed and implemented a production-grade Stripe Connect integration supporting platform commissions, delayed payouts, and transaction-safe escrow-style flows between participants.

• Architected and implemented a full Stripe Connect marketplace integration (Separate Charges & Transfers model);
• Implemented Application Fee handling for platform revenue capture;
• Designed payout orchestration using Stripe Transfers with delayed release after event completion;
• Built secure Stripe webhook verification using stripe.Webhook.construct_event for signature validation;
• Implemented idempotent webhook processing to ensure retry-safe event handling;
• Designed atomic database transactions to prevent race conditions and double transfers;
• Built a secure multi-tenant backend with RBAC and tenant isolation;
• Optimized PostgreSQL transactional consistency for payment flows;
• Led backend architecture decisions for payment infrastructure;
• Implemented document-based storage for location and event metadata alongside relational systems;
• Designed query patterns for efficient retrieval of geo-related data.

A professional social networking platform designed for skilled trades professionals, allowing users to create profiles, share content, and connect with peers through a mobile-first interface. The system included social feed functionality, support for user-generated content, and built-in monetization features such as payments and advertising infrastructure.

• Led backend development for payment and monetization features;
• Designed and implemented in-app payment flows and subscription logic;
• Integrated payment providers and ensured secure transaction handling;
• Built advertising infrastructure and monetization workflows;
• Designed APIs supporting sponsored content and ad placement logic;
• Worked on user profile systems and core social interaction features;
• Contributed to scalable backend architecture supporting user-generated content.