Every team needs an AI employee
Plan, manage and improve humans and agents the same way, in every department. And my answer to a question from GenAI Berlin: if every company runs on agents, who actually benefits?
Read articleSasa Fajkovic · Engineering leader · Speaker · Builder
I lead engineering teams and build systems that scale.
As Applied AI Lead and Staff Engineer at Trade Republic, I set technical direction, grow teams and stay hands-on with architecture and software delivery.
Berlin, Germany
Practical writing about AI agents, engineering leadership and improving how organisations work.
All articles
02
Run your agents on the hardware you already wrote off
Agents run on old, fully managed machines. Per machine the electricity costs about as much as a coffee a month. Comparable AWS capacity costs $240 to $475.
03
You cannot automate what you never wrote down
AMOS is the orchestration layer I said I would build in 2026. Nine months in, the constraint is not agents, models or protocols. It is that the processes we want to hand over have never been written down in a form anything could follow.
04
A skill is not a markdown file
Everyone is writing agent skills right now, and most of them quietly make the agent worse. What that actually costs you, why "be thorough" is the wrong instinct, and the checks that have been gating every skill across an engineering organisation for months.
I speak about applied AI, agentic systems and engineering leadership, using examples from building and operating real systems.
Next conference
9–10 November 2026Time: TBA · Workflow demo
More upcoming conferences
19 November 2026
3–4 December 2026
Previous talks
4 November 202611:45 CET · Fireside chat
Frankfurt, Germany
Banking Exchange Conference
From Internal AI Orchestration to Customer Value: Driving ROI at Trade Republic16 October 202611:15–12:00 CEST
30 September 202614:10–14:30 CEST
9 July 202615:30–16:00 CEST
Berlin, Germany
WeAreDevelopers World Congress
Agentic employees in world's most downloaded FinTech appPlanning a conference? Tell me about your audience and the problem you want the session to help them solve.
Connect on LinkedInEngineering leadership shaped by building products, teaching people and working across technology and business.
01
2026 — Present
Founder
Changify · Berlin, DE
Founded Changify, an AI, process and software consultancy that helps organisations ship production software they can own long-term.
02
2022 -> Present
Applied AI Lead, Staff Engineer
Trade Republic · Berlin, DE
Led AI adoption as well as many product engineering and platform engineering teams.
03
2020 — 2022
Engineering Manager
Deutsche Kreditbank AG · Berlin, DE
Led a cross-functional backend, frontend and QA team that delivered new onboarding flows and provisioned more than 5 million customers to a new debit card.
04
Apr 2016 — Jan 2020
Lecturer
Algebra · Zagreb, Croatia
Taught ~500 people across programming and productivity courses, including web development with PHP, HTML, CSS and MySQL.
05
Jan 2016 — Jun 2019
Engineering Manager
UniCredit Group · Zagreb, Croatia
Led a full-stack engineering team delivering core retail banking products and built CRM tooling with Do Not Call registry integration.
Full professional background
Seven roles across applied AI, banking, product building and teaching.
View every roleTechnical skills
8 areas covering backend systems, delivery, infrastructure and data.
Explore technical skillsPublications and education
Nine practical books plus a master’s degree and bachelor’s degree.
See books and educationUseful tools built from problems I encountered, documented so other engineers can inspect, run and improve them.
Agent tooling
agents-skill-eval
Evaluate agent skill packages from a local folder or GitHub URL with deterministic checks and optional AI review. Use the hosted app or run it yourself in Docker.
Unraid plugin
unraid-rsync
Schedule and monitor multiple rsync backup jobs from the Unraid web UI, with dry-runs, history and native notifications.
Personal finance
family-wealth-tracker
An open-source project for tracking family wealth. Explore the repository for its current implementation and capabilities.
Self-hosting is how I keep learning.
I run an Unraid-based home lab for storage, backups, home automation and private cloud services. It’s where I test reliable self-hosting, remote access and automation on infrastructure I operate myself.
UnraidTailscaleHome AssistantNextcloudImmichn8nPrometheusGrafanaVaultwardenrsync
Storage & private cloud
I combine on-prem and cloud solutions for my stable and experimental projects.
Home automation
Fully local, nicely automated and effective home automation using Home Assistant.
Remote access & backups
Tailscale, a separate backup node and automated rsync jobs keep it dependable.
Workflow automation
I’m building reusable workflows with n8n, with many more automations planned across my systems.