Nick Doulos

nick.os 5.0 · Safe Mode · plain-text manual

Nick Doulos

Developer with an MSc in Big Data & Analytics, Athens, Greece.

I build data systems, open-source tools and the occasional 3D web app. Most of my work so far has been about turning noisy real-world data into something reliable: GPS trajectories, streaming pipelines and the analytics on top of them.

work
FTTH Network Design Engineer at PPC S.A.
building
RouteRewind, Pyehicle and a rotating set of side projects
open to
new ideas, collaborations and things I haven't built yet
See my workEmail meLinkedIn ↗Boot nick.os

§1 Release history

Every role so far shipped as a new version.

  1. v5.0

    FTTH Network Design Engineer · PPC S.A.

    Sep 2026 – now

    • Designing fibre-to-the-home networks at PPC in Athens
    • Shipped RouteRewind: GPS tracks rendered as cinematic 3D fly-throughs, entirely in the browser
    • Certified Microsoft Power BI Data Analyst and Google AI
  2. v4.0

    Analyst & Software Engineer · Hellenic Army Informatics Support Center

    Mar 2025 – Mar 2026

    • Simulation & War Games Department: Python and R analysis for 5 computer-assisted exercises
    • Led a team of 4 that delivered three 40-minute 3D-animated training videos
    • Off-hours security monitoring and incident response under ISO 27001:2022
    • Visual design, logos and 3D modelling with open-source tools
  3. v3.1

    MSc Big Data & Analytics · University of Piraeus

    Oct 2023 – Oct 2025

    • Graduated with 9.13/10
    • Thesis: Accurate and Scalable GPS Route Reconstruction with Low Sampling Rate
  4. v3.0

    Pyehicle · Open source · GitHub and PyPI

    Jan 2024 – now

    • Grown out of my EMERALDS research and my BSc and MSc theses; first public release, 0.1.0, on 29 Nov 2025
    • HMM map-matching and 4D Kalman filtering with EM learning and RTS smoothing
    • Pandas and Polars support, Numba hot paths, chunked processing past 100K points
    • MIT licensed, documented in its GitHub README, supports Python 3.8 to 3.14
  5. v2.0

    Researcher & Software Developer · University of Piraeus Research Center

    Oct 2023 – Jan 2025

    • Horizon Europe EMERALDS: mobility data processing at the computing continuum
    • Map-matching for sparse GPS: 88.81% better F1 and 62.82% lower distance error than the HMM baseline
    • Map-matching over Apache Kafka: 4.48× consumer speedup at 5 consumers
    • Co-authored deliverable D3.1
  6. v1.0

    BSc Digital Systems · University of Piraeus

    Oct 2019 – Oct 2023

    • Major in Software and Data Systems, minor in Data Management
    • First blog post: a Sudoku solver for Project Euler problem 96
    • Thesis: Pyehicle, a Python library for preprocessing vehicle data
    • IEEE student member; volunteered at NTUA UNIque Days 2022

§2 Projects

since 2024 · active · ★ 4

Pyehicle

GPS trajectory reconstruction for Python

Turns noisy, sparse GPS points into accurate, road-aligned trajectories. 88.81% higher F1 than the standard HMM baseline on sparse GPS, from the EMERALDS research it grew out of.

PythonPolarsNumbaHMM
Open ↗Live demoRead the story

since 2026 · active

RouteRewind

Your GPS track as a cinematic 3D recap

Renders a route as a photorealistic 3D fly-through video, frame by frame, in the browser. No backend, no accounts, your data stays on your device.

CesiumJSWebGLVideo encoding
Open ↗Live demoRead the story

since 2022 · idle

WalletQR

Every QR and barcode in one wallet

Android app to scan, import or create QR codes and barcodes and keep them close for quick access or sharing.

AndroidJava
Open ↗

since 2022 · idle · ★ 2

GreatNotes

A cosy home for your notes

Dockerized full-stack web application for writing and organising notes.

DockerFull-stack
Open ↗

…and more in progress

Most of what I work on lives in private repos until it's ready, and I'm always starting something new.

§3 Writing

Subscribe by RSS

§4 GitHub

repositories
6 (4 private)
commits this year
631
contributions, last year
932
public repos
3
stars
6
followers
12
pyehicle installs, 30 days
8

Updated from GitHub every day; last public push to pyehicle 2 months ago. Languages: HTML 27%, JavaScript 25%, Python 25%, GDScript 19%, GDShader 3%, Java 2%.

Toolbox: python, sql, pandas · polars, apache-kafka, machine-learning, javascript · webgl, power-bi, r, docker, 3d-modelling.

§5 Contact

Thanks for reading this far. I'm always happy to talk about data, maps, side projects or new work.

The quickest way to reach me is contact@nickdoulos.com.

Nick

GitHub ↗LinkedIn ↗Kaggle ↗PyPI ↗