Hi, my name is
Marian Helcl Øverli.
I build intelligent systems.
Applied physicist from NTNU, Trondheim, and now diving deep into AI engineering. I'm building tools that can create real value while also having fun!
Hi, my name is
Applied physicist from NTNU, Trondheim, and now diving deep into AI engineering. I'm building tools that can create real value while also having fun!
A bit about who I am and what I do.
I'm an aspiring AI engineer with a strong interest in machine learning, natural language processing, and computer vision. I enjoy turning ideas into working code and building end-to-end ML pipelines.
I'm constantly learning and experimenting with new architectures, frameworks, and techniques. My goal is to contribute to AI systems that solve real-world problems.
Masters in applied physics from NTNU, Trondheim, 5 years experience as an acoustic consultant. Am now all in on AI Engineering
Things I've built and experiments I've run.
A conversational journaling app where you chat with Claude about your day. It summarizes the conversation into structured diary entries with themes, mood, and patterns — saved to Notion.
A privacy-focused CLI tool that records voice notes and transcribes them locally using Whisper — no cloud uploads. Supports multilingual input and saves entries to Notion.
Detect anomalies in audio using spectrograms and a CNN. Bridges acoustics domain expertise with deep learning for industrial fault detection.
Feed in arXiv PDFs and get structured summaries of key findings, methods, and results using an LLM chunking pipeline.
Samples keyframes from video, uses a vision model to understand each scene, and generates natural language narration with automatic subtitles.
Webcam-based object detection and tracking with YOLO and DeepSORT, featuring a live dashboard with stats and trajectories.
Record speech in one language, transcribe with Whisper, translate via LLM, and synthesize audio output. Built on the Voice Diary Transcriber foundation.
Technologies and tools I work with.
Thoughts on AI, learning, and building things.
Nine categories where small teams can build real AI businesses — from vertical SaaS and consulting to fintech and developer tools, with honest assessments of competition and defensibility.
Read more →What the numbers actually say about AI's energy consumption, CO2 footprint, and water usage — plus practical steps developers can take to reduce the impact.
Read more →Prompt injection, data leakage, API abuse, XSS, and privacy — the security checklist for embedding an LLM chat on your site, drawn from building my own AI Twin.
Read more →Prompt injection, deepfakes, model theft, data poisoning, agent autonomy risks, and supply chain attacks — the AI security landscape as of early 2026.
Read more →A practical overview of Supabase — its components, common use cases, the security risks developers overlook, and how to prevent them.
Read more →A deep dive into retrieval-augmented generation — how it works, when to use it, and lessons learned building my own RAG chatbot.
Read more →Practical techniques for fine-tuning large language models on limited hardware — LoRA, QLoRA, quantization, and smart dataset choices.
Read more →How I got started with AI, the resources that helped most, and advice for others beginning the same journey.
Read more →Auto-updated headlines from across the AI world.
Ask me anything — my AI twin knows about my background, skills, and projects.
This is a Gradio-powered chatbot trained on my background, projects, and expertise. Ask it anything you'd ask me in an interview!
Gradio app will be embedded here once deployed.
I'm always open to discussing AI projects, collaboration opportunities, or just chatting about interesting ideas. Feel free to reach out!
Say Hello