visitor@thomas-portfolio:~$

Pravinos Thomas

Software engineer based in Thessaloniki, Greece. I build backend systems, AI-powered developer tools, and full-stack applications, mostly with Java, Python, Spring Boot, and React. Currently at Deloitte's Engineering, AI & Data team, returning August 2026.

PythonJavaSpring BootLLMsFastAPIReactNext.jsTypeScriptC++

view projects github linkedin

// about

I grew up in Alexandria, a small city in northern Greece, and moved to Thessaloniki at 18 to study Electrical and Computer Engineering at AUTH. Six years later I'd finished a thesis on nanosatellite fault detection and landed at Deloitte building AI tools and backend systems. I'm 26, still here, and right now finishing military service in the Hellenic Army before heading back to Deloitte in August 2026.

Away from screens I follow football more closely than is strictly necessary, play video games when I can, and keep a watchlist that grows faster than I can finish it. I travel when the opportunity comes up, but most of the time I'd rather be with friends and family.

I run a home server, self-host tools I build, and keep side projects going for things my day job doesn't cover. In this way I try to discover new technologies and improve my skills.

// currently

▸ completing military service — back at Deloitte, Aug 2026

// at a glance

based in: Thessaloniki, Greece
degree: Integrated MEng, Electrical & Computer Engineering
university: Aristotle University of Thessaloniki
stack: Java · Python · Spring Boot · Next.js · LLMs

// languages

Greek

English

German

// education & research

[ AUTH ]

Integrated Master's degree — Electrical & Computer Engineering

Aristotle University of Thessaloniki

Oct 2018 – Sep 2024Grade: 7.075-year integrated Master's programme

// thesis

Fault Detection, Isolation, and Recovery (FDIR) for Nanosatellite Subsystems

This thesis focuses on the implementation of a Fault Detection, Isolation, and Remediation (FDIR) architecture based on the European ECSS Packet Utilization Standard. The C++ implementation includes the development of the parameter monitoring service ST[12], which allows for defining, enabling, and disabling checks to monitor various on-board parameters. The architecture of the service is fully configurable, allowing for the addition, modification, or removal of checks according to mission needs. Additionally, the thesis provides a detailed description of the code structure of the ST[12] service, the main methods it employs, and the logic behind the various types of checks. The proposed FDIR system aims to enhance the reliability and success probabilities of the AcubeSAT mission conducted by the students of Aristotle University of Thessaloniki.

SpaceDot · AcubeSATJul 2023 – Jul 2024 · volunteer

Implemented as part of the SpaceDot / AcubeSAT mission by the students of Aristotle University of Thessaloniki. This work was conducted as a volunteer role from July 2023 to July 2024.

↗ GitHub Repository ↗ IKEE — AUTH Library

C++ECSS StandardsEmbedded SystemsFault DetectionCubeSat

// academic projects

Telecommunication Electronics

↳ led to thesis & SpaceDot

Aristotle University of Thessaloniki

2023

Bibliographical research on how temperature affects electronic circuits in space environments. This work sparked an interest in space engineering and RF communications — the direct path that led to the thesis and SpaceDot.

↗ view repo

Graph Theory

Aristotle University of Thessaloniki

2023

Built a network of ~20,000 scientific articles/nodes linked by embedding-based semantic similarity and applied Louvain community detection to infer subject areas from titles and abstracts. Evaluated alignment with ground-truth labels using the Fowlkes–Mallows score.

↗ view repo

Radio Communications

Aristotle University of Thessaloniki

2024

Assignment for Special Topics in Propagation and Radiocommunication regarding the digital TV radio coverage analysis for the Lesvos prefecture.

↗ view repo

Optimization Techniques

Aristotle University of Thessaloniki

2023

Implementation of classical and metaheuristic optimisation algorithms for engineering problems — gradient methods, genetic algorithms, and constraint satisfaction.

↗ view repo

Distributed Production Systems

Aristotle University of Thessaloniki

2023

Bibliographical work on the role of AI in smart grids — energy management and optimisation — completed for the course Distributed Production.

↗ view repo

Computational Intelligence

Aristotle University of Thessaloniki

2022

Implemented neural networks, fuzzy logic systems, and evolutionary algorithms, an early foundation for later AI and machine learning work.

↗ view repo

// other highlights

▸11th and 12th annual ECE Student Conference of Greece·AUTH, 2019 and 2021
▸Open Workshop: AI in Energy·AUTH, Mar 2023
▸Python Programming Course·GreekLUG, May 2023

// experience

Work History

Deloitte

Software Engineer · Business Analyst

Aug 2024 – Nov 2025 · Aug 2026 – present · Thessaloniki, Greece

Working in the Engineering, AI & Data department, building smart, scalable software solutions using Java, Python, and modern technologies. Responsibilities span developing RESTful APIs, integrating databases, and applying AI to improve code quality, system performance, and team productivity.

Legacy Code Modernisation — Built a Python + LLM tool that translates legacy systems into current technologies, improving migration efficiency during client engagements
AI Documentation Platform — Developed a Python-based internal tool using LLMs to auto-generate clear documentation for complex codebases, accelerating team onboarding
Banking Services — Built secure Spring Boot microservices with REST API and database integrations supporting core internal financial operations

Hellenic Army

Research & Informatics Soldier

Nov 2025 – Aug 2026 · Veroia, Greece

Mandatory military service assigned to Research & Informatics (Special Duties). Applied technical background to support IT infrastructure and internal digital workflows.

IT systems support, data management, and internal workflow automation for military operations
Technical documentation, process standardisation, and troubleshooting
Built FireRiskMaps — a fire risk visualisation tool deployed for real internal use within the unit (see Projects)

Synapsecom Telecoms S.A.

Junior Software Engineer

Feb 2024 – Jul 2024 · Thessaloniki, Greece

Software design and development for internal applications, working across the full stack with Laravel, JavaScript, and relational databases.

Designed and developed internal applications using Laravel, JavaScript, and relational databases
Implemented RESTful APIs for application integration across internal systems
Applied Scrum methodologies, promoting agile practices and timely delivery

SpaceDot — AcubeSAT

Software Engineer (Volunteer)

Jul 2023 – Jul 2024 · Thessaloniki, Greece

Worked with the SpaceDot team at Aristotle University of Thessaloniki on software development and testing for the AcubeSAT nanosatellite — a CubeSat mission developed under ECSS aerospace engineering standards.

Developed fault-detection software for the AcubeSAT nanosatellite in C++ following ECSS aerospace engineering standards
Contributed to subsystem reliability through rigorous testing and standard compliance
Thesis: Fault Detection, Isolation, and Recovery (FDIR) for nanosatellite subsystems

MyCompany Projects

Full Stack Engineer (Internship)

Jul 2022 – Sep 2022 · Thessaloniki, Greece

Internship contributing to VCLAVIS, a Laravel/Vue.js platform for pressure vessel feasibility assessment in industrial settings.

Developed features to evaluate manufacturing feasibility of pressure vessels based on location and environmental factors
Worked across the full stack with Laravel (backend) and Vue.js (frontend)

// projects

Things I've Built

Vault

A full-stack personal finance platform built as production-ready SaaS. Track spending, income, budgets, and financial goals in one place. Users sign in with JWT authentication and interact with a Spring Boot REST API backed by PostgreSQL, while a Next.js frontend handles dashboards, forms, and day-to-day money management.

Split across vault-api and vault-frontend: the Java/Spring Boot backend covers auth, accounts, transactions, categories, and goal tracking with Spring Security and JWT; the Next.js client consumes REST endpoints for balances, trends, and goal progress. Designed with clear API boundaries, relational data modelling, and a deployable backend/frontend split typical of real SaaS products.

Spring BootJavaNext.jsTypeScriptPostgreSQLJWT

↗ api repo ↗ frontend repo

elelem

Self-hosted LLM chat app running entirely on personal hardware. No cloud APIs, no token costs, no data leaves the network. Features SSE streaming, persistent chat history, multi-model support via Ollama, intelligent model memory management, and private access over Tailscale.

FastAPINext.jsPythonTypeScriptOllamaDockerTailscaleSQLite

↗ view repo

DevTutor AI

Local AI coding tutor for programming beginners. Runs entirely on-device via LM Studio — fully private, no internet required. Interactive lessons and code explanations powered by local inference.

PythonStreamlitLM StudioOllama

↗ view repo

This site

Terminal-themed developer portfolio with an embedded AI chat widget powered by Groq. Ask it anything about my background — it answers from structured CV context streamed in real time.

Next.jsGroqllama-3.3-70bTailwindFramer Motion

↗ view repo

// built during military service

⊕ Hellenic Army · Research & Informatics · 2025–2026

FireRiskMaps

Fire risk mapping tool built during military service for actual use in the office I served at. Visualises fire risk zones and data relevant to military operational planning.

Built and deployed for real internal use.

PythonMappingData Visualisation

↗ view repo

// certifications