Art Morozov

CI/CD reliability, build systems, developer tooling, backend development

• Eliminated deployment failures caused by Jira rate limits by adding a request interceptor, Prometheus metrics, and jittered exponential backoff; drove 429-related failures to 0 across 10k+ weekly requests.
• Designed a secure CI/CD orchestration system for an air-gapped build environment by bridging GitHub events into an Kafka-like message bus and coordinating workflows across Kubernetes clusters behind the firewall and AWS, enabling PR-based testing and improving pipeline reliability.
• Redesigned Docker builds with BuildKit registry caching, cache mounts, and uv dependency locking; cut average build time from 58 minutes to 3 minutes (95%).
• Built an interactive release CLI to automate multi-repo hotfix coordination and version selection, reducing manual errors and accelerating reliable patch delivery.
• Reduced CPU and memory spikes in a critical microservice by replacing a long REST request chain with a GraphQL query that fetched only needed fields, cutting end-to-end latency by 80% and collapsing 113 downstream calls into 1.

Data pipelines, forecasting, web scraping

• Built a Python alerting pipeline using MongoDB and time-series forecasting to predict flight uploads for 100+ aerospace companies, reducing downtime by 31%.
• Developed a multi-threaded scraping pipeline with WebSocket streaming and MySQL storage, boosting throughput by 63% and enabling real-time monitoring.
• Profiled and added caching to an open-source NLP service, packaging it as an AWS Lambda API; achieved 58% performance gains and simplified scaling.
• Designed retry-safe ingestion and alerting workflows to stabilize long-running crawls during traffic spikes.

• Re-architected DynamoDB access patterns by redesigning table keys and query paths to fully leverage DynamoDB indexing, replacing sequential scans with indexed queries and cutting request latency by 67%.
• Standardized authentication email templates across Gmail, Outlook, mobile, desktop, and web by replacing unsupported SVG rendering with VML fallbacks and embedded images, delivering consistent responsive UX for 10k users.

• Built an internal monitoring portal that wraps Kubernetes health and telemetry via GraphQL Mesh, improving visibility for 1,000+ engineers.
• Implemented a CI/CD pipeline with Docker and Playwright, boosting integration and E2E test coverage by 87%.
• Migrated legacy UI components to internal React systems, improving cross-browser compatibility and maintainability.

• Implemented server-side anonymization for MRI image processing by automatically stripping sensitive patient data during ingestion, enabling secure sharing and reducing compliance risk for the platform.
• Implemented a new data model for an AI system predicting osteoarthritis diagnoses, improving accuracy by 28%.
• Built a secure MVP web platform for anonymized MRI image sharing in under 2 months, replacing manual transfers and driving researcher adoption that led to new client referrals.

• B.Sc. in Computer Science (GPA 3.8/4.0).
• Dean's List and Carleton Academic Scholarship recipient.