Harry P.

Cloud & AI Engineer | Python Developer | Agentic App Development I help teams take AI and agentic application ideas from "interesting demo" to reliable production systems. My background is in software and cloud engineering, with extensive experience building applications in Java, Node.js, and Python. Over the past five years, I've shifted my focus to cloud infrastructure architecture, managing Google Cloud environments for multiple clients. The recent evolution in my career has been extending that into generative AI and agentic application work — integrating LLMs directly into production pipelines rather than treating them as a bolt-on feature. What I bring: - LLM integration & agentic systems — Hands-on experience building agent-based operational workflows using Anthropic's models, including prompt design for structured, reliable output, tool/function-calling logic, and the surrounding code that connects model output to real actions. - Production GenAI pipelines — Integrated Gemini into live production systems, handling context management, structured output validation, batching, and error handling at the API level. - Red-teaming & reliability — I stress-test LLM-based systems for failure modes: silent inaccuracies, hallucinations, and edge cases that don't throw errors but quietly produce wrong output. - Cloud & data infrastructure — Deep, daily experience with BigQuery, Firestore, and Elasticsearch, allowing me to connect AI systems to real databases, APIs, and business logic — not just prototype in isolation. - End-to-end thinking — Comfortable working the full flow: user request → prompt/LLM → tools/APIs → retrieval or database layer → business logic → final response. I have a background in Computer Science/Engineering and approach AI work the same way I approach infrastructure: rigorously, with an eye for where it will break in production and how to catch that early.