Ibtesam Ansari

Led cross-functional teams delivering 50+ programming datasets for Google Gemini training with 100% on-time delivery. Managed Cursor Human Data Collection and RLHF workflows, including PR reviews and multi-file engineering tasks. Improved HCAST benchmarks and built ICE evaluation datasets using Python, Matplotlib, ASE, and RDKit for programmatic image generation.

Engineered 100+ high-complexity coding prompts for benchmarks including HumanEval and CodeEval to evaluate LLM proficiency in React. Developed advanced test cases and Gold Standard solutions for code generation and reasoning accuracy, while designing multi-turn scenarios targeting edge cases in React state management and asynchronous operations.

Optimized frontend architecture for Nestbox.ai by improving maintainability and reducing technical debt through modular code structures. Led technical knowledge-sharing sessions, drove adoption of TypeScript and React best practices, collaborated with clients on scalable solutions, and handled Technical Lead responsibilities including PR reviews, mentoring, and engineer training.

Improved application performance by 30% through code splitting, lazy loading, and advanced optimizations. Led a React Native migration from 0.65 to 0.75 with zero downtime, architected the multi-tenant GeoVox portal system, reduced build times by 25–30% using Vite and SWC, improved reliability to 99.9% uptime, and delivered scalable React and React Native applications.

Developed high-performance React and React Native applications, including a drag-and-drop UI engine for HRVirtuoso, the Parent Pass MVP, and the Fix Car Sharer carpooling app with real-time tracking and chat. Engineered secure identity verification flows, integrated scalable REST APIs with 99.9% uptime, and conducted R&D for frontend technologies and proof-of-concept solutions.

Architected real-time concurrency lock and session management systems for the CargoBarn logistics platform using SignalR and WebSocket-driven timers, preventing data conflicts and ensuring workflow integrity. Reduced code duplication by 40% through reusable UI strategies and led a weeklong knowledge transfer initiative with technical workshops and architectural documentation.

Architected and built JMTrophies, a high-performance e-commerce platform using Next.js 13 and TypeScript. Developed a custom engraving engine with Fabric.js (HTML5 Canvas) and integrated scalable APIs to synchronize customization with order fulfillment.

Grade: CGPA 3.823 / 4

Bachelor of Science in Computer Science at NED University of Engineering and Technology (Karachi). Achieved CGPA 3.823/4.