Olympia Dartsi

About

I started my scientific journey studying physics at the Aristotle University of Thessaloniki. Shortly after completing my M.Sc. degree in computational physics, I moved to the picturesque alpine town of Annecy, France, where I did my PhD in experimental particle physics at CERN. If you'd like to know more, you can find my PhD thesis here.

After finishing my PhD, I made the transition into data science, where I continued applying scientific reasoning to real-world problems. The topics were different, but the core mindset remained the same: combine theory with data to make sense of complex systems and uncover hidden patterns.

Currently, I'm a Fellow Postdoctoral Researcher at Lawrence Berkeley National Laboratory, working with A-Lab, an autonomous laboratory for materials discovery. My work focuses on combining chemistry, data science, and automation to make sense of complex synthesis and characterization data — blending statistical reasoning, machine learning, and interactive tools.

I revel in the problem-solving aspect, but I don't want to stop there: being able to communicate my findings and their theoretical underpinnings to audiences of wildly different backgrounds has always been dear to my heart.

"To learn is a natural pleasure, not confined to philosophers, but common to all men." — Aristotle

• Work Experience

• Education

Research

My research sits at the intersection of AI, scientific reasoning, and automation — from particle physics at CERN to autonomous materials discovery at Berkeley Lab.

Featured Work

Automated Interpretation Framework (AIF)

An open-source LLM-guided framework for automating chemical reasoning in high-throughput phase identification. Combines probabilistic methods with large language models to interpret X-ray diffraction data from autonomous synthesis experiments at LBNL's A-Lab.

Selected for Breakthrough News Oral Presentation at MRS Spring Meeting 2026.

LLMs Chemical Reasoning Autonomous Labs XRD Python

Interactive Dashboard for Materials Research

Designed and deployed a unified data-interactive dashboard integrating synthesis and characterization data streams into a single queryable interface for the A-Lab autonomous laboratory.

Dash/Plotly HoloViz Panel Data Integration Python

PhD: Particle Physics at CERN

Developed boosted decision tree algorithms for signal selection within petabyte-scale datasets in the ATLAS experiment. Implemented Bayesian unfolding techniques to correct for detector effects, contributing to large-scale collaborative physics analysis at the LHC.

Read the thesis →

BDT Bayesian Unfolding ATLAS/LHC C++ Python

Revenue Forecasting & Campaign Optimization

At Tempr, predicted future revenues for advertising campaigns via neural network-based time series forecasting. Optimized budget allocation, audience, and time targeting using multi-armed bandit algorithms and k-means clustering.

Time Series Neural Networks Multi-Armed Bandit K-Means

Code & Experiments

Anomaly Detection (CV)

Unsupervised anomaly detection in MNIST digits with ROC AUC evaluation.

Parkinson UPDRS Forecasting

Predicting MDS-UPDRS scores to measure Parkinson's disease progression.

Time Series (ARIMA)

Time series forecasting using the ARIMA model.

Multi-Armed Bandit

UCB algorithm implementation for exploration-exploitation tradeoffs.

Data Imputation

Evaluating imputation methods for handling missing values in time series.

Google Foobar Challenge

Complete solutions for all levels of Google's secret programming challenge.

• More on GitHub

Publications

Papers

See all publications on Google Scholar.

Selected Talks

Outreach

Knowledge transfer is one of the key aspects of science. I've always been drawn to simplifying and visualizing complex concepts for diverse audiences.

Lectures & Presentations

AI Fundamentals

Introduction to artificial intelligence for general audiences.

Neural Networks 101

Visual introduction to neural networks and deep learning.

Data Visualisation Techniques

Basics and advanced techniques for visualizing scientific data.

From Academia to Data Science

Mindset and tools when transitioning from research to industry.

PhD Thesis Defense

Summary of research contributions during my PhD at CERN.

Writing

No More Scrum for Data Scientists

Reflections on how agile methods like Scrum may not fit scientific workflows.

The Ambiguity of Charisma

Exploring how we define and misinterpret charisma in leadership.

Mentorship

Scientific Mentorship: Mentored a research intern at LBNL on a project evaluating the reliability and safety of LLM prompting strategies (OpenAI vs. open-source models).

Community Mentorship: Providing guidance to high school students in Greece on careers in STEM and Data Science through The Tipping Point.

Life

When I'm not sifting through bits and bytes, you'll find me somewhere outdoors — running mountain trails, cycling across continents, or painting. I also hold a certificate in Braille, because why not.

Adventures

More on my ITRA runner profile and Strava.

Paintings

Painting is my preferred form of meditation — here are a few selected works.

What I'm reading: Goodreads.