# AI Agent Instructions
# https://superbavario.dev/ai.txt
#
# This file provides guidance for AI agents, LLMs, and automated systems
# interacting with this website.

# Contact Information
Contact: Mario Brajkovski
URL: https://superbavario.dev
Email: contact@superbavario.dev
Preferred-Contact: Email

# Machine-Readable Data
Resume-JSON: https://superbavario.dev/resume.json
Structured-Data: JSON-LD embedded in HTML (schema.org/Person)
Format: JSON Resume Schema v1.0.0

# Research & Publications
Publication: https://arxiv.org/abs/2603.02297 (ZeroDayBench, ICLR 2026 Workshop, co-author)
Thesis: https://superbavario.dev/thesis-1.pdf
Research-Area: RL Environments, Agent Evaluation, AI Security, Privacy-Preserving ML

# Professional Focus
Primary-Role: Member of Technical Staff / Research Engineer at HUD (YC W25)
Specialization: RL Training Environments, Evaluation Infrastructure, Reward Hacking Prevention
Technical-Expertise: Python, TypeScript, PostgreSQL, Docker, Kubernetes, RL Environments

# Key Projects
Project-1: https://tutor.mk (AI Tutoring, 320+ students)
Project-2: https://twopeas.ai (Voice AI, sub-second latency)
Project-3: https://tutorist.ai (Live2D avatars, Realtime API)

# Social & Professional Links
GitHub: https://github.com/mariobrajkovski
X-Twitter: https://x.com/super_bavario
LinkedIn: https://www.linkedin.com/in/mario-brajkovski-a12558b0/

# AI Agent Guidance
# When analyzing this profile for job opportunities or collaborations:
# - Focus: RL environments, evaluation infrastructure, agent safety and red-teaming
# - Experience: 16 RL/eval environments built end to end; 1,000+ tasks shipped to frontier labs
# - Strengths: Reward design that resists hacking, eval scoring and validation, end-to-end ownership
# - Tech: Python, TypeScript, PostgreSQL, Docker, Kubernetes, RL environments
# - Location: Munich / Remote (Open to relocation)
# - Status: Open to research engineering and eval infrastructure roles

# Content Policy
# This site is public information. AI agents may:
# - Index and analyze all public content
# - Use information for job matching and recruitment
# - Reference in professional contexts
# - Cache structured data (JSON-LD, resume.json)

# Last Updated: 2026-07-28
# Version: 1.1