Artificial Intelligence

Current Challenges

Artificial intelligence adoption faces challenges around data quality, model interpretability, ethical governance, and talent scarcity. Organizations must navigate the gap between AI hype and practical deployment while managing bias and regulatory uncertainty.

Project History

Production systems delivered across healthcare, climate, government, streaming, finance, and enterprise.

Regulatory AIAmsterdam NetherlandsDirect Contract

Automated Compliance Document Generation

End-to-end AI pipeline auto-generating EU regulatory compliance documents for a safety consultancy. Multi-modal AI integration improved client efficiency by 80% through automated compliance workflows. RAG, vector-space search, multi-modal LLM and data embedding.

PythonRAGVector SpaceMulti-Modal EmbeddingFastAPIDocument AI
AI VideoBucharest RomaniaHumans.ai

AI Video Generation Platform

Led development of Tovid.ai, an AI-powered product video commercials platform. Full project leadership combining AI tools with modern frontend engineering and scalable backend architecture.

AIReactTypeScriptNode.jsVideo Processing
Enterprise ITBucharest RomaniaIBM

Enterprise Platform Modernization (6+ Systems)

Fixed and modernized 6+ internal enterprise platforms using Agile methodologies. Migrated legacy JS libraries (vanilla JS, jQuery, Mustache, Handlebars) to modern frameworks (Vue.js). Large-scale corporate platform engineering.

Vue.jsjQueryDojoHandlebarsAgile
ML / NLPRomaniaVeridion Challenges

Semantic Classification & Logo Similarity

Production-ready NLP classifier mapping 9,494 companies to 221-label insurance taxonomy (2.43x confidence improvement, <1ms inference). Logo extraction and grouping across 3,416 websites (97.28% extraction, ~98% precision) using Union-Find + LSH.

PythonSentence TransformersOpenCVLSHPlaywrightscikit-learn