AI Engineer · Agentic Systems · Interpretability
ANDREI
AI ENGINEERING
I build and study agentic systems: the harnesses, tools, evaluations, and recovery paths that make modern models usable in production.
Before AI engineering, I spent 13+ years in evidence-based human-facing work. Today this site is focused on AI: agentic workflows, interpretability, RAG, reasoning, and the software layer around modern models.
01
BACKGROUND
2022 – 2025
AI & Sustainable Technologies, BSc
Tomorrow University, Germany
Current
AI systems practice
Agentic workflows, RAG, interpretability, and reasoning
Previous
Evidence-based professional background
13+ years in structured human-facing work before AI engineering
02
FOCUS
01Agentic workflows that recover from tool failures and partial execution
02Evaluation harnesses, observability, and permission models for AI systems
03RAG and agentic search patterns for investigative, multi-hop questions
04Interpretability and verification practices for production AI teams
03
PRACTICE
AI engineeringAgentic systems, RAG, interpretability, and reasoning workflows
Applied researchTurning emerging model capabilities into reliable product behavior
Evidence-based practiceA prior evidence-based background that shaped the operating style, not the service offering
WritingNotes on agent infrastructure, AI reliability, search, and societal impact
Building, studying, and writing about agentic systems.