Zilan Aze Karabay

– Led the development of a smart home system leveraging Deep Learning (LSTM) models for high-accuracy energy price forecasting and subsequent real-time optimization. – Implemented Reinforcement Learning (RL) agents to proactively manage and control home appliances, achieving significant energy efficiency improvements. – Containerized the entire application using Docker and managed deployment via Kubernetes (K8s) on a Cloud Platform, ensuring system reliability and scalability.

– Contributing to Generative AI projects focusing on Large Language Models (LLMs) and advanced automation workflows. – Designed and deployed a RAG (Retrieval-Augmented Generation) Proof of Concept using Make (Integromat) for automated knowledge retrieval and contextual response generation. – Built and optimized LLM-based APIs for interactive applications, integrating with LangChain, OpenAI API, and Vector Databases (FAISS / Pinecone). – Developed internal tools for prompt engineering, model evaluation, and response consistency analysis using Python and FastAPI. – Tech Stack: Python, LangChain, OpenAI API, Make (Integromat), FastAPI, RAG, FAISS, Pinecone, LLMs, Prompt Engineering, API Orchestration, Automation

– Developed and fine-tuned Large Language Models (LLMs) to enhance predictive intelligence systems. – Built end-to-end AI applications integrating LLM APIs, vector databases (Pinecone, FAISS), and FastAPI. – Designed prompt-response optimization pipelines improving model accuracy and reliability. – Collaborated with cross-functional teams using React, Node.js, and Docker for scalable deployment. – Tech Stack: Python, FastAPI, React, Node.js, Docker, LLM, RAG, Prompt Engineering, API Integration, PostgreSQL, Machine Learning