Bio
Ben Bykovski is a 19-year-old AI and automation engineer based in Amsterdam, currently in his second year of an Honours Computer Science degree at Eindhoven University of Technology (TU/e), where he ranks in the top 10% of his cohort. He builds real production systems — not demos — that automate document-heavy work inside legal, real estate, and finance teams. Since March 2026, he has been an AI automation intern at a major European tech company, where his flagship project is a regulatory-intelligence platform used daily by a ~70-person legal team: it scrapes regulatory updates, runs automated parliamentary trackers, and delivers self-updating briefings via hybrid SQL and vector retrieval.
Ben is the founder of BYK, an AI and automation studio targeting document-heavy teams with a fixed-price, no-slideshows model. His ethos is explicitly anti-hype — BYK positions itself against consultancies selling strategy decks at €50k and agencies that never ship. On the hardware side, his Honours project is a burnout detection device for ER nurses using custom ESP32 hardware with PPG/HRV sensing, designed to catch chronic stress before it becomes breakdown. The analytics layer is live; AI-driven detection is next.
Beyond client and academic work, Ben is building Kovos — a personal AI operating system for research, memory, and daily automations, his real-world answer to the 'personal Jarvis' problem. The name comes from the Slavic root 'kov-', meaning to forge. He competes too: he took first place at AI Fixathon Amsterdam in the Work Rights track, building an accessibility-first tribunal navigator for vulnerable workers in 24 hours. His stack is RAG, pgvector, Supabase, Flask, and Next.js.
What I'm building
Real products that automate work and working products that save time for companies using AI (in Legal, real-estate, finance, etc.), building a personal Jarvis (my 24/7 assistant that I give everything to), and cool things like stress detection devices that can track and predict burnout
Why TAG
Found out through Tijs' Linkedin, following him since the Fixathon at Vandebron last year that me and my friend won 1st place at, and was super fun and a great experience with amazing people! Love the initiative of TAG and want to be part of this community of cool people and builders!
Projects
AI and automation studio offering fixed-price builds, audits, and workshops for document-heavy teams in legal, real estate, and finance. Explicitly positioned against consultancies that sell strategy decks instead of shipping.
Personal AI operating system handling research, memory, and daily automations via Telegram — Ben's long-running 'personal Jarvis' project. Named from the Slavic root for 'to forge.'
Production RAG system used daily by a ~70-person legal team: scrapes regulatory updates, runs automated parliamentary trackers, and delivers self-updating briefings via hybrid SQL and vector retrieval.
Custom ESP32 hardware with PPG/HRV sensing designed to detect and predict chronic stress in ER nurses before it escalates to burnout. Analytics layer live; AI detection in progress.
Accessibility-first web app for vulnerable workers navigating UK employment tribunals, built in 24 hours at AI Fixathon Amsterdam. Supports 8 languages with a streaming AI chatbot.