I'm passionate about building clean, efficient software and data pipelines, especially in NLP and full-stack web development.
Profile-based, deterministic text curation pipelines designed for large-scale NLP datasets.
- Structured, versioned pipelines for reproducible data preprocessing
- Used for preparing corpora in LLM training and evaluation
- Focus on semantic preservation, deterministic transformations, and scalability
- Integrated tightly with Hugging Face Datasets ecosystem
- Written in Python, Apache 2.0 licensed
A modern chat application built with Next.js and TypeScript.
- Real-time messaging with clean UI
- Backend and frontend fully integrated
- Designed for ease of deployment on Vercel
- Tech stack: Next.js, TypeScript, Python, CSS
