Contract Software Engineer, AI Model Trainer at Outlier (Scale AI)

Review hundreds of AI-generated code outputs weekly, providing structured feedback to improve coding accuracy for millions of end users. Lead reviews and collaborate with global teams of 10–20 engineers to maintain quality and align with client goals. Trained 20+ LLMs using Java, Python, TypeScript/JavaScript, C/C++, and SQL while applying internal metrics to prioritize language- and category-specific weaknesses. Built testing suites with 200+ test cases and standardized rubrics to ensure consistent evaluation across projects. Accelerated model failure identification by 30% through robust testing suites. Improved overall model training efficiency by 20% by leveraging internal performance metrics. Designed and delivered 100+ production-level coding tasks exposing weaknesses in LLM code generation, debugging, and refactoring, leading to measurable accuracy gains. Authored onboarding guides and process documentation that reduced new contributor ramp-up time by 30%.

Full-stack Developer at Tango

Built a Kanban project management platform for 50+ students with Java Spring Boot, React, and Bootstrap. Implemented GitLab CI/CD pipelines, cutting deployment time by 50% and post-release bugs by 20%. Integrated Spring Security for authentication and role-based access in Agile sprints that halved update cycles. Documented system architecture and UI/UX with style guidelines, ER diagrams, README, and Figma designs.

Full-stack Developer at Creator 2023

Developed and launched a full-stack anime recommendation platform with Spring Boot and React. Built a recommendation algorithm using web scraping, Jikan API, and collaborative filtering on MyAnimeList data. Optimized performance with asynchronous API calls, throttling, and stream processing, cutting API requests by ~90%. Applied advanced data structures (Priority Queue, Hash Set, Hash Map, Concurrent Collections) to rank and aggregate recommendations efficiently.