Mindaugas K

Worked on AI-heavy projects involving data cleaning, filtering, and training models with AMD GPU using PyTorch and DirectML. Created a prototype model to distinguish melanoma-linked and healthy skin spots. Performed accessibility and performance audits improving Core Web Vitals scores to above 95 and boosted user retention by 20%. Provided SEO audits and optimization, contributing to improved user traffic. Set up a Gitlab CI pipeline for testing, deployment, and notifications via Slack. Resolved server log storage issues, improving uptime from weekly crashes to 99.95%. Utilized Google Cloud Platform Translation API to translate JSON documents into over 32 languages.

Built cloud infrastructure including firewalls, reverse proxies, API gateways, and event bus installation. Developed REST API endpoints and WebSocket servers using Nest.js. Fixed Gitlab CI pipelines and deployed releases on Kubernetes. Created secure authentication for non-public resources, collaborated on complex business logic issues, and wrote unit/integration tests. Implemented integrations with Ethereum blockchain and Apache Kafka event bus for microservice data replication and event handling in distributed systems.

Analyzed and maintained large backend (Node.js, Express.js) and frontend (React, Redux, JQuery) codebases. Managed huge SQL databases (>20TB) and developed real-time data processing applications on Node.js. Monitored errors using Kibana and Zabbix, handled unpredictable data reliably. Worked with web scrapers, crawlers, proxies, database engines, and microservices brokers. Engaged in emergency production bug fixing including Docker and Kubernetes-related processes. Ensured accurate score/stats/odds data for sports gambling to guarantee correct payouts. Used Jira and version control systems for collaboration.