My Resume

Golang implementation of the SymSpell algorithm for ultra-fast spelling correction

• Developed a memory-efficient algorithm for spelling correction and fuzzy search
• Implemented support for both unigrams and bigrams for advanced contextual correction
• Optimized for performance, achieving correction speeds up to 1 million times faster than traditional approaches

Persian Text Normalizer for search optimization

• Created a customizable Persian text normalizer for cleaning and preprocessing text
• Implemented tools to handle spaces, URLs, punctuation, and other text elements
• Designed for integration with search engines to improve Persian language search quality

Go-Based Backend Service Framework

• Developed a robust and flexible foundation for building backend services in Go
• Automated essential tasks to simplify the process of creating backend systems
• Designed for both experienced developers and newcomers to Go backend development

Snapp, Tehran, Iran

• Refactored and re-architected core services, reducing resource usage by 50% and improving system scalability
• Redesigned business logic for driver-passenger matching, achieving a 15%+ increase in driver acceptance rate, directly boosting key business metrics
• Collaborated cross-functionally with other product teams to enhance feature integration, resulting in a measurable increase in user adoption of the carpooling service
• Partnered with business stakeholders to accelerate feature delivery, helping the team meet critical deadlines 50% faster than company expectations

Snapp, Tehran, Iran

• Engineered and deployed AI-powered search and serving models using Seldon Core, enhancing real-time search efficiency and improved Search Metrics By 5%
• Integrated a high-speed Spell Corrector for the search engine using SymSpell, reducing query errors and handling millions of requests in milliseconds and led to improve Search Metrics by 3%
• Operated a highly available map autocomplete system with Golang, Elasticsearch, and Redis, processing 17 million daily requests and significantly reducing search latency
• Orchestrated the development of a reverse geocoding service with Golang, Envoy, and Tile38, handling 5,000+ requests per second and improving address accuracy by 5%, leading to better user engagement
• Designed and automated a self-healing data update pipeline, leveraging Python, Ansible, k3s and Kubernetes. Reduced update time by 50%, eliminated manual tasks, and saved 1,000+ engineering hours annually
• Enhanced search accuracy by integrating a sentence-to-vector model, expanding query coverage and boosting relevance by 10%
• Implemented a metric collection service capable of handling 400 RPS, storing data in ClickHouse, which led to a 12% increase in user engagement via real-time analytics
• Developed a training pipeline for a Named Entity Recognition (NER) model, improving query relevance and accelerating booking speed by 10%
• Designed an online search history feature for seamless device synchronization, enabling personalized search experiences

HICH (Human Intelligence Community Hub), Birmingham, United Kingdom (Remote)

• Directed the migration from a monolithic architecture to microservices, scaling system capacity to handle 2,000 RPS across multiple services, improving reliability and fault tolerance
• Revamped the Notification service with an event-driven architecture, enhancing real-time message delivery and reducing system bottlenecks
• Managed a cross-functional team in rewriting a legacy PHP project into Golang, improving performance by 3x, reducing resource usage by 30%, and increasing maintainability

Carriot, Tehran, Iran

• Constructed and maintained microservices using Golang, applying principles of Clean Code, Domain-Driven Design (DDD), and Hexagonal Architecture, reducing development time by more than 10%
• Utilized gRPC for communication and integrated an event-based system through RabbitMQ, handling over 1k events per second for seamless inter-service event management
• Employed Postgres and Clickhouse for database management, with Redis for caching and Elasticsearch for full-text search, optimizing performance and handling over 100 requests per second
• Conceptualized and built a Internal Single Sign-On (SSO) service, fortifying security with real-time access level validation, achieving a 5 ms maximum response time
• Boosted development speed by 20% by optimizing workflows and reducing bottlenecks