Aashay Kulkarni

Background

About Aashay Kulkarni

I'm passionate about building AI systems that actually work in the real world. Currently working as a Machine Learning Engineer at KOBIL. My journey started with curiosity about how machines could learn independently. This led me to pursue my Master's in Computer Science at Saarland University, where I spent over two years as a researcher at the August-Wilhelm Scheer Institute. During this time, I worked on energy forecasting using neural networks, built federated learning systems that could train models without sharing sensitive data, and explored AI applications for mental health challenges through my thesis on cognitive distortions. My supervisors often mentioned my ability to "quickly master unfamiliar topics"- something that's served me well in the rapidly evolving world of generative AI, where prompt engineering techniques change constantly. This research foundation now drives my work at KOBIL, where I apply the same curiosity and technical rigor to enterprise AI challenges. These days, I'm excited about agentic AI and building systems that can reason and act autonomously. I work with different LLMs, craft effective prompts, and build RAG systems that help organizations make sense of their data. I also love federated learning and retrieval systems - there's something elegant about distributed training while preserving privacy. Core Skills: AI/ML Frameworks: PyTorch, TensorFlow, HuggingFace, Scikit-learn, Keras Generative AI: LLM fine-tuning, prompt engineering, RAG systems, agentic AI Specialized ML: Federated learning (Flower), neural networks (CNN, RNN, LSTM), anomaly detection ML Security: Membership inference attacks, AI safety protocols, model evaluation Data & Infrastructure: Python, R, SQL, Docker, MongoDB, Git Research: Published papers, energy forecasting, Graph Neural Networks, NLP applications If you're working on interesting AI challenges, I'd love to connect: aashayk7@gmail.com