Work

On a mission to turn messy data into decisive insight and to bring cutting‑edge LLM research out of the lab into everyday operations.

ropa GmbH

I joined ropa — a fast-growing German fibre-optic provider — just as the company realised spreadsheets could no longer keep pace with its network expansion. My goal was straightforward: build a reliable, single source of truth that each department could trust.

I began by defining clear KPIs tied to real-world challenges faced by field operations, network control (NOC), strategic building planners, customer service, and the C-level team — ensuring visibility into customer connections, identifying network issues, supporting strategic expansion decisions, and pinpointing customer service trends. Introducing Jira and Confluence streamlined requirements management and task tracking, helping me translate scattered stakeholder needs into manageable, prioritised tasks. In taking on these responsibilities, I naturally stepped into the role of product owner for ropa’s analytics platform.

From there, I architected a PostgreSQL-based data warehouse using a medallion structure (bronze → silver → gold), layered with dbt for transformations, and migrated pipeline orchestration from Airflow to Dagster — making nightly ELT workflows as predictable as cron jobs. Seven disconnected systems — ranging from CMS entries to network infrastructure APIs (covering router provisioning, service activation, and monitoring) and public geodata — now flow smoothly through Airbyte and Python transforms, feeding nine curated data marts (~1M fresh rows nightly).

With reliable data established, I shifted attention to visibility. Power BI dashboards provided clarity: executives now access reports refreshed in less than three minutes (down from overnight waits), field technicians view real-time fibre fault updates on tablets, and customer service teams proactively spot trending support issues. Alongside technical delivery, I mentored analysts unfamiliar with Git-based workflows, set coding standards, and communicated regularly with management, turning their feedback into actionable steps.

Today, ropa’s analytics platform underpins daily operational decisions and strategic planning, fostering a culture where data genuinely guides business growth rather than being an afterthought.

Public Sector

A national audit and consulting firm for hospitals and public institutions — asked me to untangle its reporting backlog. I joined a team of experienced data engineers to built an in‑house MSSQL warehouse that stitches together ERP, HR, CRM, and project‑accounting feeds, then surfaced the numbers through self‑serve Power BI models. A few well‑placed indexes, table partitions, and rescheduled ETL windows cut median query times from twelve seconds to four and pushed job reliability past 99 percent.

But the real win came from automation. By packaging finance and compliance checks into Python micro‑services running on FastAPI and RabbitMQ, I freed consultants from forty hours of monthly spreadsheet drudgery. Working closely with the software department, I also exposed clean data to other systems via C# REST endpoints and, when a needed dataset didn’t exist, spun up large‑scale web‑scraping jobs to create it. Six months later the advisory teams could answer clients’ questions in minutes, not days, and the firm had its first glimpse of what a modern data stack could deliver.

Earlier Work

Before stepping into data leadership roles, I built up a diverse engineering foundation across startups, universities, and enterprise clients.

Pexon Consulting At Pexon Consulting, I worked closely with enterprise teams as a cloud consultant, advising on AWS and Azure architectures and improving CI/CD pipelines. Despite a short tenure, I delivered several small-scale end‑to‑end projects under tight deadlines and was frequently praised for autonomy and fast ramp-up.

University Student Project During my undergrad time at Solent University, I led a cross-functional student development team building machine learning–enabled backend services related to students mental wellbeing. We used Python, GraphQL, SQL, and Docker to ship a production‑ready prototype under real-world constraints. It was my first experience managing both people and pipelines — and it stuck.

42 Ads & aveta As a full-stack developer at 42 Ads, I took ownership of several Vue.js and Node.js projects, balancing frontend polish with backend reliability. At aveta, I got hands-on with PHP and custom frameworks, contributing to a specialized CMS and learning to move fast without breaking everything.

Internship at Rohde & Schwarz and proISMS Earlier on, I immersed myself in mobile development with Kotlin and Android Studio — spending three months at Rohde & Schwarz and four months at proISMS GmbH, where I delivered working prototypes, debugged platform-specific issues, and learned how to ship solid code even as an intern.

Each of these roles sharpened a different edge — whether building scrappy MVPs, automating workflows, or translating stakeholder needs into software. Together, they laid the groundwork for my transition into modern data engineering and AI‑driven solutions.

Beyond Work

Outside of my day job, I’m passionate about using technology to create access and opportunity. For over a year, I served as the volunteer technical lead at bisonaide, I’ve spearheaded the development of Whakaako, an open education platform aimed at making IT training freely available across the globe. From infrastructure design to full-stack implementation, I’ve helped shape a tool that lowers the barrier to entry for learners regardless of background or location.

I also enjoy sharing what I learn. Whether it’s speaking at international conferences like DSAA 2025 in Birmingham or ICBCT 2025 in Muscat, I focus on translating complex research — like LLM-driven topic modeling or retrieval-augmented generation — into practical strategies for teams looking to integrate AI into real-world workflows.

These side projects and speaking opportunities keep me grounded in both the human and technical sides of innovation. They remind me that code is most powerful when it’s shared — and when it solves problems that matter.

🛠️ Skills Snapshot

Data Engineering

PostgreSQL · MSSQL · dbt · Dagster · Airbyte · Power BI

Backend & APIs

Python (3yrs) · FastAPI · C#/.NET · RabbitMQ

Cloud & DevOps

Docker · GitHub Actions · Azure · AWS · Linux

AI & NLP<

Transformers · LLMs · HuggingFace · spaCy · OpenAI APIs · Qdrant · ChromaDB

Workflow & Project Delivery

Agile & Scrum · Stakeholder Communication · Team Leadership · Public Speaking