Oussama Dadouch

Full-Stack Developer

Vilnius, Lithuania

I build the internal tools companies do not know they are missing until they have one.

00.About

I work as a Level 2 Technical Solution Specialist at a fintech international company based in Vilnius, Lithuania. I'm a payment incident and technical issue solver for online and onsite payments.

Alongside that role, I design and build AI-powered internal tools on my own initiative, things like retrieval systems for support knowledge and anomaly detection for transaction data. I like finding the slow, manual part of a process and turning it into something that runs on its own.

01.Experience

  1. Level 2 Technical Solution Specialist @ MarketPay

    June 2025 - Present

    Handle complex technical escalations for a live payments platform. Alongside that core role, I designed and shipped two AI-powered internal tools now used daily by the support team: a knowledge retrieval assistant and a real-time transaction anomaly detection system.

  2. Customer Success Specialist @ Hostinger

    December 2022 - November 2023

    Helped hosting customers work through technical and account issues over chat and email, everything from DNS and email setup to server and billing problems, in a fast, high-volume support queue.

02.Projects

The code for these projects is under NDA, so I cannot share repositories publicly. I am happy to walk through the architecture and the decisions behind it on request or in an interview.

AI Knowledge Retrieval Assistant

NDA

Problem: Support agents kept losing hard-won troubleshooting knowledge in a graveyard of old Zendesk tickets, and re-solving problems that had already been solved once.

What I built: A retrieval-augmented generation system on Google Cloud Vertex AI. It indexes daily Zendesk ticket exports along with Confluence and Jira content, so the team can ask a question in plain language instead of digging through search results.

Impact: Cuts down the time agents spend hunting for a fix that someone on the team already found before.

  • Google Cloud Platform
  • Vertex AI
  • RAG
  • Python
  • REST API

OMA, Open Market Alerting System

NDA

Problem: At the volume this business processes, unusual transaction activity could slip by unnoticed until it was already a problem.

What I built: A Python system that checks 5,000 to 10,000 transaction events every 5 minutes against a database of millions of rows, and flags anything that looks off in real time. I paired it with a dashboard for filtering results by client, merchant, or transaction ID.

Impact: Anomalies get flagged within minutes instead of turning up days later in a manual review.

  • Python
  • SQL
  • REST API
  • Next.js

Grafana Dashboard Automation Scraper

NDA

Problem: A vendor quoted 5,000 euros a month for an API to pull our dashboard data, and until then four people were checking dashboards by hand every day.

What I built: A browser automation scraper built with JavaScript and Puppeteer, written in my first month on the job.

Impact: Replaced the 5,000 euro a month vendor proposal and eliminated the manual dashboard checks four people used to do.

  • JavaScript
  • Node.js
  • Puppeteer
  • Grafana

03.Contact

Want to talk about a project, a role, or how any of this actually works under the hood? My inbox is open and I read everything.

oussamadadouch@gmail.comgithub.com/ELMudyr
Download resume

Loosely designed in Figma and coded in Neovim by yours truly. Built with Next.js and Tailwind CSS, deployed with Vercel. All text is set in the Inter typeface.