I build LLM-enabled web apps with clean, scalable architectures, fast intuitive interfaces, and reliable data pipelines—focused on measurable performance, cost efficiency, and long-term maintainability.
Framework-agnostic TypeScript charting for financial and big-data workloads, rendering through WebGL2 with a plugin-first kernel and streaming support.
Headless, accessible color picker primitives across 15 color spaces and four frameworks, with a benchmarked zero-dependency color engine and GPU-rendered gradients.
AI-powered semiconductor news analytics platform with news aggregation, sentiment charts, a RAG chatbot and daily summaries. Built from scratch as a solo full-stack developer.
Quant signal execution platform that combines model outputs with Reddit-based sentiment, running on Prefect ETL flows, LLM sentiment analysis, and cost-efficient ClickHouse data pipelines.
All-in-one CRM and booking platform for extracurricular education businesses. Built with Clean Architecture and Feature-Sliced Design, keeping business logic first so a startup can change its mind.
Sentiment analytics platform for traders. It aggregates social media posts, predicts sentiment with ML, and visualizes the result through charts, treemaps and graph networks. My first commercial project, and the one where I grew from frontend to full-stack developer.