Johan Lahti

Background

About Johan Lahti

I have a background in GIS and spatial data science combined with over fifteen years in software engineering and analytics. The last year, I have been working full-time with AI features on a variety of tasks – ranging from toy exercises to developing production ready SaaS services – such as- RAG "agent": embedding PDF:s into a vector database with a reranker model at the end - Fine-tuning LLM:s with LoRA - Augmenting LLM:s with knowledge graphs - Embedding knowledge graphs for compatibility with LLM:s - Model hosting of embedding and reranker models using Docker and Kubernetes - Developing an API in Go lang - Developing ML algorithms for anomaly detection, in Python - Developing frontend visualisations in React and Typescript Overall, I have easy to move between disciplines and I thrive at the intersection different fields, such as data science, software engineering as well as other emerging AI/ML fields such as Causal AI, Operations Research (OR), reinforcement learning (RL) and symbolic AI. If you have similar interests or would like to discuss opportunities, feel free to reach out to me. And if you are curious of what I am coding, you can find some of my work on GitHub: https://github.com/johanlahti