hello, I'm

Viktor Sharha

A student at the University of Edinburgh

Developer Portfolio

Developer Portfolio

October 2025
Next.js
Tailwind CSS

Modern, performant single-page portfolio built with Next.js and Tailwind CSS. Implements adaptive theming with system preference detection and smooth scroll-anchored navigation. Optimized component architecture ensures fast load times and seamless user experience across all devices.

GradeIQ

GradeIQ

October 2025 (2-Week Sprint)
Next.js
FastAPI
Supabase
shadcn/ui

Third-place Teachathon winner: AI-powered grading platform automating academic assessment. Engineered responsive Next.js frontend with TanStack Query for real-time polling and optimistic updates; built FastAPI backend with non-blocking background tasks; secured with Supabase JWT authentication. Integrates Claude and Gemini APIs for intelligent assignment evaluation.

Samantha OS Assistant

Samantha OS Assistant

October 2025
Python
OpenAI API
Linux

First-place winner in university OpenEuler Challenge (10 competing teams): intelligent CLI assistant that converts natural language into system commands. Chains complex operations across file management, diagnostics, and web research with safety confirmations for destructive actions. Built on extensible agent framework with 5+ specialized tools and automatic retry mechanisms.

Wordle International

Wordle International

September 2025 -- October 2025
React
Redux Toolkit
FastAPI
Tailwind CSS

Feature-rich multilingual Wordle supporting 15 languages with configurable game modes. Built FastAPI word service with offline fallback; implemented Redux Toolkit state management for guess validation and keyboard interactions; designed responsive Tailwind interface with locale-aware input and touch-friendly controls. Supports URL-parameterized game state for easy sharing.

Words-FrequencyPOS

Words-FrequencyPOS

September 2025
Python
Bash

Cross-lingual frequency + POS dataset pipeline powering language-learning and NLP research. Automated ETL that merges multilingual word-frequency and POS datasets, links frequency data with morphology metadata, and produces reproducible, versioned artifacts with idempotent downloads.