Hi, I'm Evgeni.
Cluj, Romania
Forward-deployed Solution Architect and AI consultant at Tecknoworks. I help companies find where AI makes a real difference, then design and deliver it through to production, across domains such as legal tech, manufacturing, mining, insurance and others.
I also run AI-First SDLC workshops for engineering teams and speak at conferences like DevTalks Cluj.
Here I write about AI-first engineering, agentic AI and AI-led business transformation, with case studies from real engagements.
Selected work
All case studies-
Operating model
The AI-First SDLC
The approach 8 of our 12 engineering teams now run on, presented at DevTalks Cluj.
-
Case study
Business Transformation: 31 Pain Points to AI in Production
A two-week assessment into a funded roadmap, and roughly 80% less manual finance work in production.
-
Personal system
A Second Brain with Obsidian + Claude Code
An AI-maintained knowledge base in Obsidian, with Claude Code as the structuring engine.
Latest writing
All posts-
The AI-First SDLC: From Coding to Context Engineering
A practical guide to the AI-First SDLC. AI made writing code fast, but capturing intent is still slow: here's the operating model we run on, with a Project Brain, Living Specs and Agentic Loops, and the lessons we learned along the way.
-
The AI Use-Case Handbook: 16 Demos of What Actually Works
The AI use-cases we keep seeing across industries that actually deliver business value in production — natural-language BI, document extraction, contract review, churn, forecasting, computer vision and more. Sixteen reusable patterns, each with a short visual demo.
-
Spec-Driven Development: A Practical Guide for AI-Accelerated Teams
A comprehensive guide to Spec-Driven Development: why specs are the missing layer for AI-accelerated teams, and how to put it into practice.
-
How I Built My Second Brain with Obsidian + Claude Code
A practical guide to building an AI-native personal knowledge base in Obsidian, with Claude Code as the structuring engine. From manual Notion to automated cross-domain wiki in an afternoon.
-
Multi-Agent AI with Hybrid Search: Cutting Document Review Time by 80%
How a multi-agent orchestrator combined with LanceDB's hybrid search cut document review time by 80% — practical patterns from a Legal Tech production build, with full source code.
-
DevOps for Azure: Multi-Environment Provisioning With Terraform & CI/CD Pipelines
Step-by-step guide to setting up a multi-environment DevOps solution with CI/CD pipelines and Azure resource provisioning using Terraform and Azure DevOps.
Working on an AI initiative that needs to reach production?
I'm happy to compare notes, whether you're scoping a first use case or trying to get a stalled pilot over the line.