Full-stack developer Available full-time
I buildsoftware peopleactually use.
Final-year Computer Engineering student and full-stack developer working across React, TypeScript, .NET and SQL — from polished interfaces to APIs and production debugging.
Theofanis TompolisNicosia, CyprusIntegrated Master's · Expected 2027
{ build: ship }
REACT — TYPESCRIPT — .NET — SQL — SIGNALR — DOCKER — LINUX — TESTING — REACT — TYPESCRIPT — .NET — SQL — SIGNALR — DOCKER — LINUX — TESTING —
01 / Selected work
Things I've built
and worked on.
01Production full-stack workPlatform closed · Case study
Footy Greece
Contributed to a multi-application sports platform with React 19, TypeScript and ASP.NET Core. My work covered registration, role-aware chat, attendance reporting and data consistency. The platform has since closed; the case study documents my contributions.
React 19TypeScriptASP.NET CoreSignalRSQL
02Data engineering / full-stackPublic source
Biblio Explorer
A bibliographic analytics platform combining Python ETL, Flask REST APIs and MySQL with a JavaScript/D3.js dashboard. Integrates DBLP publications and venue rankings for search, author profiles and interactive comparisons.
PythonFlaskMySQLJavaScriptD3.jspytest
Data engineering / full-stack↗
Publication records processed2.5M+
ETLClean & match venues
APIProfiles & search
D3.jsInteractive comparisons
Explore the work ↗03Java / application engineeringPublic source
Traineeship Management
A Java 17 and Spring Boot application for traineeship management, with authentication, role-based workflows, placement matching, logbooks and evaluations. Controller and service behavior is covered with JUnit 5 and Mockito.
Java 17Spring BootMySQLJUnit 5Mockito
Java / application engineering↗
Placement → logbook → evaluationROLE-AWARE
AuthRole-based access
MatchPlacement strategies
TestControllers & services
Explore the work ↗04Machine learning / NLPPublic source & report
Food Hazard Detection
Text classification for SemEval-2025 Task 9: predicting food hazard and product categories from recall reports. Compared word/character TF-IDF and LinearSVC pipelines with a DistilBERT baseline; the final reported Subtask 1 score was 0.78003.
Pythonscikit-learnTF-IDFLinearSVCNLP
Machine learning / NLP↗
Reported final Subtask 1 score0.78003
10Hazard categories
22Product categories
SVCFinal classifier
Explore the work ↗05Compiler designPublic source
Greek++ Compiler
A multi-stage compiler for a Greek-language programming language. Python lexical and recursive-descent analysis lead to intermediate quadruples, symbol tables and RISC-V assembly, with support for functions, procedures and parameter passing.
PythonLexerParserQuadruplesRISC-V
Compiler design↗
A complete compilation pipelineSOURCE → RISC-V
.grGreek++ source
.intIntermediate code
.asmRISC-V assembly
Explore the work ↗02 / Experience
Production work,
not just coursework.
03 / Stack
Tools I actually
work with.
04 / About
Across the stack,
not stuck in one layer.
I like understanding how the whole product works — interface, API, data and the messy debugging in between.