Savvas Mantzouranidis

Constructed and optimized distributed data scraping and manipulation pipelines using Python multithreading to automatically populate vector databases, reducing manual developer operations by 30%. Developed a retrieval augmented generation (RAG) back-end API with Python, FastAPI, and LangChain, improving semantic search and AI conversational agents, enhancing search accuracy by 25%. Built an AI-featured front-end UI with SvelteKit, Node.js, and TypeScript including semantic search, visualizations, and an AI chat agent. Led decentralised machine-to-machine communication interface using Python, FastAPI, and WebSockets for real-time data streaming. Deployed and maintained cloud-based serverless solutions on AWS (S3, ECS Fargate, DynamoDB, etc.), reducing energy consumption by 12%. Automated CI/CD pipelines using GitHub Actions. Integrated Redis caching to reduce API response latency by 40%. Optimized container deployment with multi-stage Dockerfiles, reducing image size by 60%. Composed He

Constructed and optimized distributed data scraping and manipulation pipelines with Python multithreading, reducing manual developer work by 30%. Developed a RAG back-end API using Python, FastAPI, and LangChain to enhance semantic search and AI conversational agents, improving search accuracy by 25%. Built AI-featured front-end UI using SvelteKit, Node.js, and TypeScript, including semantic search, data visualization, and AI chat. Led development of decentralized machine-to-machine communication with Python, FastAPI, and WebSockets for secure real-time streaming. Managed cloud serverless deployments with AWS services reducing on-premises energy consumption by 12%. Wrote CI/CD pipelines with GitHub Actions for automatic deployment, integrated Redis caching to reduce API latency by 40%, and optimized Docker containers to reduce image size by 60%. Authored Helm charts for Kubernetes deployments and collaborated cross-functionally for technical alignment and timely project delivery.

Integrated an open-source authorization agent in a TypeScript-based IoT monitoring edge service to generate and inspect access tokens for sensors, increasing system security by 70%. Developed a Python multithreaded device emulator simulating hundreds of virtual devices to enable extensive API stress and penetration testing, enhancing system security and robustness.