Bartlomiej Tedys

Hey there! I'm Bartek (because "Bartlomiej" breaks most keyboards). I'm a 21-year-old developer from Poland but grew up in Ireland building websites, analysing data, and teaching AI to recognise patterns in everything from customer behaviour to satellite imagery.

The Origin Story

Started coding when I was 12 because I wanted to make video games. Turns out game development is hard, but web development and data analysis? Those clicked immediately. Now I'm the person businesses call when they need both a website that converts AND insights from their data that actually help them make better decisions.

What I Actually Do

Build websites and web apps that convert visitors into customers (recent client saw 300% more leads)
Create data dashboards that turn spreadsheet chaos into clear business insights
Develop AI/ML models for everything from computer vision to predictive analytics
Design data pipelines that automatically collect, clean, and analyse your business data
Build React and Next.js applications with integrated analytics and reporting features
Transform raw data into visualisations that even your least data-savvy team members can understand
Write Python scripts that automate boring data tasks so you can focus on strategy
Translate between "what the data says" and "what this means for your business"

The Technical Arsenal

Web side: JavaScript/TypeScript, React, Next.js, Tailwind CSS, and whatever gets your site fast and functional. Data side: Python, pandas, PyTorch, TensorFlow, SQL, and all the tools needed to make sense of your data. I also speak fluent "can you make this into a chart?" and "what does this data actually mean?"

Why This Combo Matters

Most developers build websites. Most data analysts crunch numbers. I do both, which means I can build you a website that not only looks great but also tracks the right metrics, provides actionable insights, and helps you understand what's actually driving your business growth. It's web development with a brain for data.

Munster Technological University, Cork, Ireland

2021 - 2025 (just finished!)

Bachelor of Science (Honours) - Software Development

Four years of learning how to break things properly before fixing them. The course covered everything from "Hello World" to deploying apps that thousands of people actually use. I particularly enjoyed the modules where we got to build real projects instead of toy examples.

Final Year Thesis: I compared three different AI models to see which one is best at spotting individual trees in aerial photos. Turns out machines are getting scary good at this stuff.

Hogeschool van Amsterdam, Netherlands

2023 - 2024 (Erasmus Exchange)

Big Data & Mobile App Development

Spent a year in Amsterdam learning about big data while trying not to get run over by cyclists. Built a mobile app for a local startup and discovered that working with international teams means debugging code in three different time sones.

What I Actually Learned

How to build web apps that don't fall over when people use them
Machine learning beyond the hype (most problems don't need AI, but when they do, here's how)
Database design that won't make future developers hate you
Mobile development that works on both iPhone and Android (it's harder than it looks)
How to work in teams without losing your sanity
Why documentation matters (learned this the hard way)
Docker and Kubernetes (because apparently everything needs to be containerised now)
How to explain technical concepts to non-technical people (invaluable skill)

The Real Education

The best learning happened outside lectures: building actual projects, debugging at 2 AM, working with real clients, and figuring out how to make Amsterdam rent on a student budget. University taught me to code; experience taught me to solve problems.

The Challenge

January 2025 - Ongoing | Freelance Project

A local construction company was losing potential customers to competitors with better websites. Their old site looked like it was built in 2010 and worked about as well as you'd expect. They needed something fast, professional, and most importantly - something that would actually bring in leads.

What I Built

Completely rebuilt their online presence from scratch
Mobile-first design (because most people browse on their phones)
Lightning-fast loading times (nobody waits for slow websites)
Clear call-to-actions that actually guide visitors to make contact
SEO optimisation so people can actually find them on Google
Analytics tracking to measure what's working

The Results (This Is The Good Part)

Within 2 weeks of launch, they saw a 300% increase in online leads. Not page views or bounce rates or other vanity metrics - actual potential customers filling out contact forms and calling them. That's the kind of result that makes this job worth it.

Why It Worked

I kept it simple. No fancy animations or trendy design elements that distract from the main goal. Just clean, professional design that loads fast and makes it easy for visitors to understand what M&L does and how to hire them. Sometimes vanilla HTML, CSS, and JavaScript is exactly what you need.

Check It Out

See the live site at mlrenovations.ie

Project Screenshots

Homepage
Hero Section

Services
Overview

Contact
Form

Mobile
Responsive

Analytics
Dashboard

Performance
Results

Teaching Machines to Count Trees From Space

February 2024 - July 2024 | Research Internship

Joined a research project that's using AI to track deforestation in real-time. The goal? Build computer vision models that can automatically detect and count trees in satellite imagery. It's like playing "Where's Waldo" but with forests, and the stakes are saving the planet.

What I Actually Did

Built and trained deep learning models using PyTorch and TensorFlow
Processed thousands of satellite images to teach AI what trees look like from space
Optimised models to run fast enough for real-world use (turns out speed matters when you're monitoring entire forests)
Created automated pipelines that can process new satellite data without human intervention
Wrote documentation that other researchers can actually understand and use

The Cool Technical Stuff

We used state-of-the-art object detection architectures, fine-tuned them on custom datasets, and got them running efficiently enough to process real satellite feeds. The tricky part wasn't just detecting trees - it was distinguishing individual tree canopies from each other in dense forests.

Why This Matters

Traditional forest monitoring takes months and requires teams of people manually analysing images. Our AI can do the same analysis in hours and catch deforestation as it happens, not months later. It's environmental protection with a tech upgrade.

What I Learned

Working on research taught me that building AI for real-world problems is completely different from following online tutorials. You spend most of your time on data quality, edge cases, and making sure your model actually works when it matters.

Research Gallery

Model
Architecture

Satellite
Imagery

Detection
Results

Processing
Pipeline

Performance
Metrics

Production
System

My Amsterdam Adventure

September 2023 - December 2023 | Exchange Project

During my Erasmus semester in Amsterdam, I joined a small Dutch startup called Binqy to build their mobile app. It was my first time working in a real startup environment, and definitely my first time debugging React Native code while trying to avoid cyclists.

What We Built

Cross-platform mobile app using React Native (one codebase, two platforms)
Real-time features that kept users engaged without draining their batteries
Clean, intuitive interface that worked for both tech-savvy and casual users
Backend integration that actually stayed connected when users switched between WiFi and mobile data
The whole thing from concept to app store submission

The International Team Experience

Working with developers from different countries taught me that good code translates across cultures, but communication styles definitely don't. I learned to over-document everything, ask questions early, and never assume everyone is on the same page just because they said "yes" in the meeting.

Real-World Lessons

University projects have clear requirements and deadlines. Startup projects have neither. Requirements change weekly, deadlines are "as soon as possible," and sometimes you discover the feature you just spent two weeks building isn't actually what users want. It's chaotic but incredibly educational.

The Result

We shipped a working app that real people downloaded and used. Seeing actual user reviews (even the negative ones) was both terrifying and exciting. Nothing teaches you to write better code like knowing strangers are depending on it to work.

App Development Gallery

Main
Interface

User
Dashboard

Key
Features

Design
Mockups

Team
Photo

Launch
Day

Making Data Less Boring

Academic project that turned me into a data visualisation convert. The assignment was simple: take complex environmental data and make it understandable. The execution was anything but simple.

The Challenge

Raw environmental data that looked like spreadsheet soup
Multiple data sources with different formats and update schedules
Users who needed insights, not just pretty charts
Everything had to work on mobile because that's how people browse data now

My Solution

Built an interactive web platform using D3.js for custom visualisations and vanilla JavaScript for the logic. Created filtering systems that let users explore the data instead of just staring at static charts. Added real-time updates so the information stayed current.

The D3.js Learning Curve

D3.js is powerful but has the steepest learning curve I've ever encountered. I probably spent more time reading documentation than writing code. But once it clicked, I could create exactly the visualisations I wanted instead of settling for whatever the charting library offered.

What I Learned

Data visualization isn't just about making things pretty - it's about telling stories with data. The best charts aren't the most complex ones; they're the ones that help users understand something they couldn't see before.

Project Screenshots

Main
Dashboard

Interactive
Charts

Filtering
System

Mobile
View

Code
Architecture

Final
Presentation

AI Tree Detection Showdown

My final thesis project: comparing three different AI models to see which one is best at detecting individual trees in aerial photographs. Think of it as a heavyweight championship, but for computer vision algorithms that specialise in forest analysis.

The Research Question

Built a custom dataset of annotated tree images from various aerial sources (turns out manually marking thousands of trees is as tedious as it sounds)
Implemented three competing approaches: Mask R-CNN, YOLACT, and a custom YOLO-based solution
Tested each model on speed, accuracy, and how well they handle edge cases like overlapping tree canopies
Benchmarked performance across different image qualities and forest types

Making It Practical

Integrated the best-performing model with QGIS so forestry professionals could actually use it
Created automated pipelines that process raw aerial imagery and output georeferenced tree maps
Built it to work with real-world data, not just clean academic datasets

What I Discovered

The "best" model depends entirely on what you're optimising for - speed vs accuracy vs computational resources
Data quality matters more than model complexity (garbage in, garbage out, even with fancy AI)
Academic papers often skip the messy parts of making AI work in production environments

The Real Contribution

Beyond comparing models, I created a reproducible workflow that other researchers can use and improve upon. The QGIS integration means forestry professionals can use this technology without needing a computer science degree.

Thesis Project Gallery

Segmentation
Results

Dataset
Creation

Performance
Comparison

QGIS
Integration

Testing
Pipeline

Final
System

Light Mode? Really?

About Me

Education

Projects & Experience

Projects & Experience

M&L Home Builds Website

Computer Vision Research - Caeli

University Thesis Project

Binqy Mobile App

University Data Visualisation Project

Light Mode? Really?

Bartlomiej Tedys

About Me

Education

Projects & Experience

Projects & Experience

M&L Home Builds Website

Computer Vision Research - Caeli

University Thesis Project

Binqy Mobile App

University Data Visualisation Project

About Me

The Origin Story

What I Actually Do

The Technical Arsenal

Why This Combo Matters

Education

Munster Technological University, Cork, Ireland

Hogeschool van Amsterdam, Netherlands

What I Actually Learned

The Real Education

Let's Build & Analyse

Message Sent! 🎉

Oops! Something Went Wrong

M&L Home Builds Website

The Challenge

What I Built

The Results (This Is The Good Part)

Why It Worked

Check It Out

Project Screenshots

Computer Vision Research - Caeli

Teaching Machines to Count Trees From Space

What I Actually Did

The Cool Technical Stuff

Why This Matters

What I Learned

Research Gallery

Binqy Mobile App

My Amsterdam Adventure

What We Built

The International Team Experience

Real-World Lessons

The Result

App Development Gallery

Data Visualisation Project

Making Data Less Boring

The Challenge

My Solution

The D3.js Learning Curve

What I Learned

Project Screenshots

University Thesis Project

AI Tree Detection Showdown

The Research Question

Making It Practical

What I Discovered

The Real Contribution

Thesis Project Gallery