|
I craft solutions that move smoothly from notebook insight to deployable service. My work spans automated ML pipelines, GenAI prototypes, and data-driven analytics, always with an eye on maintainability and impact.
Bridging the gap between ML experimentation and real-world AI applications.
I'm an AI/ML Engineer who loves turning ideas into working intelligent systems. I get excited about the entire journey — from exploring data and training models to designing end-to-end pipelines that make those models actually useful.
What drives me is the challenge of connecting the dots — taking a messy real-world problem, finding the right data and modeling approach, and then packaging it into something people can actually use. Whether it's GenAI, classical ML, or time-series forecasting, I'm always looking for the most elegant path from question to answer.
Everything here I built end to end. The code is public, each project has a write-up on the decisions behind it, and a short demo video if you would rather watch than read.
End-to-end AI/ML projects — from data exploration to working applications.
End-to-end MLOps pipeline for electricity price forecasting across European energy markets. Features multi-horizon predictions, ensemble methods, and comprehensive model comparison with MLflow integration.
A privacy-first document processing system that transforms PDFs into conversational interfaces. Built with local AI inference, vector search, and advanced RAG capabilities for intelligent document analysis.
Production speech recognition application built with HuggingFace Transformers and the Hubert model. Features real-time audio transcription with a clean, user-friendly Streamlit interface.
Smart document parsing system with advanced image and text extraction capabilities. Supports multi-format PDFs with table question-answering for structured data retrieval.
Practical insights from building AI systems, exploring new tools, and learning in public.
Exploring MCP, the open standard that's revolutionizing how AI assistants connect with external tools and data sources. A deep dive into architecture, capabilities, and real-world applications.
Building a production speech-to-text application using HuggingFace transformers and the Hubert model. Lessons learned, performance benchmarks, and practical tips for audio ML projects.
How I built a privacy-first document assistant that turns PDFs into conversational partners. Exploring chunking strategies, embedding models, and the magic of retrieval-augmented generation.
From raw ENTSO-E data to real-time predictions — how I built a complete MLOps system for European electricity price forecasting with FastAPI, MLflow, and Cloud Run.
Have a project in mind or want to collaborate? I'd love to hear from you.
Whether you're looking for an AI/ML engineer for your team, want to discuss a project idea, or just want to say hello — my inbox is always open.
I'm always excited to work on challenging ML problems and build production-grade AI systems.