Data Scientist & Engineer
Hi, this is Sian! I love working at the intersection of data and AI — taking systems that feel complex and messy and turning them into something reliable, useful, and ready to deploy.
My work spans AI engineering, data engineering and machine learning. I build end-to-end pipelines that move from raw data through transformation, model training, and into production. Lately I have been focused on large language models: designing LLM-powered workflows, building retrieval-augmented generation systems, and exploring how AI can be applied in healthcare and business to create real impact.
Outside of work, I spend a lot of time taking care of my plants 🌿, exploring new places, and staying active. I am someone who enjoys simple things, being close to nature, and bringing a sense of calm and intention into both life and work.
Skills
LLM RAG NLP Machine Learning MLOps Python FastAPI Docker SQL PostgreSQL Spark dbt AirflowA selection of the day's work
A full-stack job search tracker supporting companies, applications, and interview rounds with status filtering, CSV bulk import, and a live dashboard. Backed by PostgreSQL with Alembic migrations, deployed on Render and Streamlit Cloud.
End-to-end data engineering pipeline aggregating live crypto market data from CoinGecko, Binance WebSocket streams, and the Fear & Greed Index. Stores data in S3 as Parquet, loads into Redshift, transforms with dbt (Kimball star schema), and visualizes in Metabase. Orchestrated by Airflow, provisioned with Terraform.
End-to-end MLOps pipeline for predicting diabetes risk. Covers the full lifecycle — data ingestion, model training, experiment tracking, API serving, containerization, CI/CD, drift monitoring, and data versioning.
Cloud-based big data project exploring behavioral trends in political engagement on social media, with NLP and ML analysis on Reddit data.
Data cleaning, text mining, and machine learning applied to a clinical trials dataset, exploring data science questions behind trial outcomes and patterns.
A weather forecasting project on climate change data using linear regression, vector autoregression, and LSTM neural networks. Published in ICJE.
Modeling healthcare company stock prices using multiple time series methods to explore development and financial trends of the sector.
more coming soon...
A little corner of calm
