I plan. I build. I review. I test. I ship.

I build teams of specialized AI agents that work alongside me, coding, testing, reviewing, researching and iterating on your product. My broad experience, amplified by AI, lets me approach your product as a whole: understand it, plan it, build it, challenge it, test it and ship it.

Book a call See the work

Currently building

In the field: 12+yrs
Engineers led: 50+
Tests shipped: 25000+
Pipeline speed-up: 775+%

Typically responds in < 24h

Šemsudin S.

AI Engineer · Dev · QA · DevOps · Product

Available for select engagements
Sarajevo, BiH · GMT+2

Trusted by

Three areas I work in.

See the work

AI Systems Architecture 01 / 03

Multi-agent GenAI platforms in production.

I design and build agents on LlamaIndex and LangChain that work with 10+ LLMs through one interface, with RAG over company documents (parsing, chunking, pgvector), streaming responses and OpenTelemetry tracing. I care about the parts that keep a system running after launch: monitoring, testing and safe upgrades.

Read case study

LlamaIndex
FastAPI
RAG
pgvector
Azure OpenAI
Multi-LLM

PDF · DOCX · PPTX · XLSX · OCR → pgvector

Proof Enterprise GenAI platform: React UI, .NET 8 orchestrator, Python agent, 10+ LLMs.

AI-Driven QA 02 / 03

AI agents that write and maintain tests.

I build AI agents that explore the UI and write test plans, turn tickets into Playwright tests, fix tests when selectors change, and move tickets along based on results. They run inside the development cycle, so regressions, OWASP top-10 issues and performance problems show up early.

Read case study

Playwright
AI Agents
ISTQB AI Testing
Self-healing
Security scans

Test run · 6 apps

1000+tests

6apps

US + UKregions

Test plan written from UI exploration plan

Playwright test created from work item tests

Selector changed, test updated self-healing

Ticket moved on test results ticket flow

OWASP top-10 scan on the PR security

- page.locator('#signup-btn')
+ page.getByRole('button', { name: 'Sign up' })

Authenticated pass rate ~~20%~~ → 100%

Proof 1000+ tests across 6 apps for two pet care brands. Authenticated test pass rate 20% → 100%.

Enterprise DevOps 03 / 03

CI/CD, infrastructure and cloud.

One pipeline handles 7 services, builds only what changed and deploys in parallel. Production images are built once and promoted between container registries. Infrastructure is written in Bicep with per-service RBAC, and test coverage from .NET and Python is reported in one place.

Read case study

Azure DevOps
GitHub Actions
Bicep
Docker
Clean Architecture

QA sign-off gate
Waiting for E2E

Proof Automation rescue: 300% more stable, 775+% faster test runs. GenAI platform: one pipeline for 7 services.

Independently verified.

Woven is an independent company that runs technical work simulations. Their engineers scored my solutions, and all four assessments came back as a strength.

Scored by

Work simulation · Oct 2026

Fullstack .NET Engineer
Strength

Senior

3 of 3 scenarios rated a strength
- Debugging Social Media
- Managing User Permissions
- Change Password Alert w/ AI
“We don't have any feedback for you on potential improvements because your work accomplished almost everything we check for. Well done!”

Woven feedback
Work simulation · Aug 2026

DevOps Engineer
Strength

Senior

3 of 4 scenarios rated a strength
- Product Recommendations are Slow
- Monolith Service Extraction on EKS
- Cleaning Up Old Files
“We liked that you specified all parts of the infrastructure we look for.”

Woven feedback
Work simulation · Jul 2026

Java CLI
Strength

2 of 3 scenarios rated a strength
- Managing User Permissions
- Cleaning Up Old Files
“The code passed all 11 tests, including a few tricky edge cases.”

Woven feedback
Work simulation · May 2026

GenAI Engineering
Strength

Senior

2 of 3 scenarios rated a strength
- Debugging Social Media
- Prorating Subscriptions
“Your solution passed 12 out of 12 tests we used to validate correctness, including 10 tests that were not shown during the work simulation.”

Woven feedback

Case studies.

All work

01 / 04

Industrial Manufacturing · Enterprise GenAI

Enterprise GenAI platform

QA Lead → Sr. SDET → AI Systems Architect

I've worked on four projects for a global industrial company over seven years. I started by modernizing their test automation and went on to work on the architecture and development of their enterprise GenAI platform and an agentic development pipeline.

Engagement: Long-term
Distinct projects: 4
Pipeline speed-up: 775+%
LLMs orchestrated: 10+

Read case study

02 / 04

MarTech · Consumer

AI-driven QA

Senior QA Engineer · AI-Driven QA Lead

I built a Playwright framework from scratch covering 6 projects across two brands of a global pet care company, in the US and UK. I also built AI agents that write test plans, create tests from work items and fix failing tests.

Automated tests: 1000+
Apps covered: 6
Auth pass rate: 20% → 100%
Regions supported: US + UK

Read case study

03 / 04

Parking Operations · Microservices

QA transformation

Lead QA Engineer · QA Transformation

I led QA on the core project of Europe's largest parking operator, a microservices platform. I set the test strategy, processes, test stages and definition of done, and built and led the QA team working inside the delivery pipeline.

Role: Lead
Architecture: Microservices
QA team: Led

Read case study

04 / 04

Enterprise Software · Regulatory Technology

Test automation architecture

Test Automation Architect · QA Strategy Lead

I worked on two projects in a row on a European regulatory and registry platform. On the Core project I set up the QA strategy and automation framework. On the Cyprus rollout I led testing across several countries and moved the framework from Playwright/TypeScript to Cypress/JavaScript.

Regression runtime: −orders of magnitude
Engagements: 2 back-to-back
Countries deployed: Multiple

Read case study

Twelve years ago I started testing software.

I'm Šemsudin (most people write it "Semsudin"), an AI Engineer based in Sarajevo. In 12+ years in IT I've always considered myself a jack of all trades, master of none, working across development, QA, automation, DevOps, infrastructure and product.

AI changes that equation.

More about me

Playbook · Free · PDF

The Enterprise GenAI Reliability Checklist.

42 checks in 6 categories that I go through when reviewing an enterprise AI system. Each one comes from a real production incident and is written as a pass/fail item you can print and use.

Get the PDF See what's inside

Leave your email and I'll send you the PDF. No newsletter.

I plan. I build. I review. I test. I ship.

Trusted by

Three areas I work in.

Multi-agent GenAI platforms in production.

AI agents that write and maintain tests.

CI/CD, infrastructure and cloud.

Independently verified.

Fullstack .NET Engineer

DevOps Engineer

Java CLI

GenAI Engineering

Case studies.

Enterprise GenAI platform

AI-driven QA

QA transformation

Test automation architecture

Twelve years ago I started testing software.

The Enterprise GenAI Reliability Checklist.

Let's ship the next one together.