Experience
Refract Labs
| Founding Engineer | Dec 2023-Present | New York, NY |
Sole engineer responsible for designing and building the entire Refract ecosystem from ground up, delivering multiple complex products with aggressive timelines while maintaining high code quality and scalability.
Refract Node Service
- Architected and implemented full-stack tokenization platform integrating Shopify with Sui blockchain, achieving sub-second responsiveness
- Engineered high-performance Express backend and React frontend with TailwindCSS, completing MVP in under 2 weeks
- Developed secure OAuth2 authentication system for Shopify merchants with robust session management
- Built scalable token redemption infrastructure capable of handling thousands of concurrent requests
- Implemented comprehensive automated testing suite achieving 95% code coverage
Refract Developer Platform
- Designed and built multi-tenant developer platform supporting diverse blockchain workflows with 50ms average API response time
- Engineered flexible tokenization engine supporting multiple protocols with plug-and-play architecture
- Implemented sophisticated CRON-based workflow engine handling time-based triggers, webhooks, and on-chain events
- Created intuitive setup wizard reducing new project onboarding time from hours to minutes
- Achieved zero downtime deployments through robust CI/CD pipeline and infrastructure automation
Refract Passport
- Architected high-throughput wallet signing engine handling 15,000+ real transactions in first week
- Implemented advanced gas optimization system reducing API response time by 60%
- Built scalable multi-wallet gas station supporting 1,500+ concurrent users with zero failed transactions
- Developed real-time notification system integrating Telegram Bot API with 99.99% delivery rate
- Engineered robust error handling and recovery systems achieving 99.9% system reliability
Cookie Shark
- Built engaging on-chain game driving 1,500+ user acquisitions in first week
- Achieved 100% user conversion to Refract Passport through seamless wallet integration
- Implemented real-time game mechanics with <100ms latency for optimal user experience
- Designed modular CI/CD architecture allowing rapid feature iteration and deployment
- Created automated analytics pipeline providing real-time insights into user behavior
Technologies: React, Next.js, Express, Node.js, Sui Blockchain, WebSocket, gRPC, Telegram Bot API, OAuth2, TailwindCSS, CI/CD, AWS
Auura
| NLP Engineer Intern | Sept 2023-Dec 2023 | Haifa, Israel |
- Architected and implemented semantic search engine achieving consistent 92% similarities to search results
- Built responsive React frontend with real-time search suggestions, reducing average search time by 40%
- Engineered high-performance Django backend capable of handling multiple concurrent searches with sub-second latency
- Worked on automated data pipeline processing 100,000+ NPO records daily
- Created comprehensive competitor analysis framework leading to 3 major feature implementations
Technologies: Huggingface, Transformers, PostgreSQL, React.js, Django, Docker, ElasticSearch
Stoopit
| Full-Stack Engineer and Project Manager | Jan 2023-May 2023 | New York, NY |
- Co-Led team of 4 engineers, delivering MVP 2 weeks ahead of deadline
- Architected scalable MERN stack application with interactive Google Maps
- Implemented automated CI/CD pipeline reducing deployment time from hours to minutes
- Practiced Agile with GitHub Project with comprehensive project tracking
- Established comprehensive testing framework achieving >95% code coverage
Technologies: React.js, Node.js, MongoDB, Express, GitHub Actions
Diacritize
| Lead Engineer | Jan 2024-May 2024 | New York University |
- Designed and implemented three distinct approaches for Arabic text vocalization: Viterbi N-gram, RNN, and Transformer models
- Built comprehensive data pipeline processing extensive Quranic texts
- Engineered custom evaluation framework to enable cross-architecture performance comparison
- Implemented efficient data preprocessing reducing training time by 40%
- Created extensible architecture allowing easy integration of new models
- Achieved diacritic error rates between 3.5-7% across different architectures
Technologies: PyTorch, RNN, HMM, Python, Docker
NLP Projects
- Part-of-Speech Tagging System with Viterbi Algorithm achieving 92.4% accuracy on WSJ Test set
- Noun Groups Sequence Labeling with MaxEnt Bio Tagger achieving 96.96% accuracy on WSJ Test set
- Information Retrieval System with TF-IDF on cranfield dataset achieving MAP score of 0.38 (score: 10/11)
Education
New York University
Sep 2021 - May 2024
Bachelor of Arts - BA, Computer Science
Bachelor of Arts - BA, Linguistics