Nikola Posarac

Designed and published two Chrome extensions from scratch — technical, design, distribution — without a team, without a budget.

Pareto Capture: transforms any YouTube video into structured notes using AI. 25+ formats available (Zettelkasten, Cornell, flashcards...), collaborative rating system on thumbnails to evaluate a video before watching, and a chat module with the summarized content.

Freemium model published on the Chrome Web Store.

ParetoSub: solves a concrete problem — automatic YouTube subtitles in French are often unusable.

The extension intercepts audio in real-time, sends segments to an AI model, and displays synchronized French subtitles.

Zero configuration, one-click activation.

What these projects demonstrate: the ability to identify a real problem, build a complete solution, and put it in the hands of users — not just prototypes.

Instead of paying €500 to €1,500/month in LinkedIn subscriptions, I built my own prospecting system — perfectly adapted to my way of working.

What makes it different: it includes an AI trained solely on my data. It learns from every interaction, every message sent, every post published.

Each week, it generates reports and suggests concrete adjustments — what types of messages convert better, which post formats perform well, where my response rate is dropping.

The more I use it, the more it refines its recommendations. I no longer rely on instinct; I make decisions based on data about my own behavior.

Result: automated pipeline, continuously optimized personalized messages, LinkedIn content tailored to my audience.

Total cost: €0.30/month.

This is the type of system I build for my clients — a custom tool, adapted to their exact process, with or without AI depending on the need.

Not a generic tool that you install and forget.

At the heart of one of Europe's largest asset management companies, I worked on two parallel dimensions: creating and monitoring investment funds in collaboration with marketing teams and fund managers, and transforming my team's operational processes through automation.

Multi-source data extraction: the team spent 15 hours per week manually consolidating data on our funds from various systems. I developed a custom Excel tool that automatically centralizes, extracts, and displays this data.

Result: 12h → 1h per week per employee, for a team of 5 people.

Automation of communications: drafting and sending communication emails to over 1,000 employees worldwide took 2 hours per week. I automated the entire process — content generation, recipient selection, sending.

Result: 2h → 10 minutes per week.

AI & agent integration: in direct collaboration with Amundi's Head of AI, I contributed to the deployment of AI agents connected to internal tools via Model Context Protocol — allowing language models to interact directly with existing data systems without heavy infrastructure.