Alice Lépissier, PhD · Machine Learning Engineer

About

Machine learning engineer and PhD-trained data scientist with 13 years of interdisciplinary experience bridging AI, climate, and global finance. Experienced in designing and deploying robust, end-to-end analytical systems using a mature, production-ready tech stack. Grounded in applied statistics and computational social science, I bring analytical rigor to both the methods and the domain. Recognized for delivering high-impact solutions and translating advanced statistical and ML methods into actionable insights for stakeholders ranging from policy-makers to technical teams. Skilled at communicating complex concepts clearly and persuasively, through high-level political advocacy (UN, UK government, NGOs), public tools and tutorials, and highly rated teaching. Trilingual and internationally raised across Taiwan, Abu Dhabi, and Spain, I bring unmatched cultural adaptability and the ability to engage with any audience.

Machine Learning Engineer & Computational Social Scientist

Currently Director at Green Finance AI
Previously Brown University, Center for Global Development, United Nations Economic Commission for Africa, Tax Justice Network, ICLEI Africa
13 years of experience
PhD Environmental Science and Management from University of California, Santa Barbara
Master's degrees Statistics, Economics and Public Policy, and Economic History
Alumna University of California, London School of Economics, Sciences Po Paris, École Polytechnique (X), University College London

PhD-level trained in statistical data science and machine learning
From data ingestion to model deployment: full-stack ML delivery
Cutting-edge in open and reproducible science
Interactive dashboards and tools
An eye for design and polished presentations
Expertise in climate finance, carbon markets, ESG, sustainable investments, illicit finance, international trade
Communicating complex concepts to lay audiences

Facts

Publications

Public Interventions

Open-Source Repositories

Awards

Students

(Human) Languages

Tech Stack

I bring quantitative depth and full-stack engineering skills to the design, deployment, and scaling of machine learning solutions. My experience spans statistical modeling, MLOps, cloud architecture, and containerized scientific computing. I've built reproducible pipelines, apps for scientific computing, interactive dashboards, and infrastructure for multi-user platforms using modern deployment stacks. Obsessed with design and data viz.

Statistical & ML Programming

R 100%

Python 99%

ML libraries: scikit-learn, XGBoost, PyTorch 90%

Generative AI stack: LangChain, Chroma, Streamlit, Hugging Face, Transformers 85%

Stata 100%

General Purpose Programming

Bash shell 90%

Linux 85%

Javascript 70%

Web & Visualization

HTML/CSS/Sass 95%

TeX 100%

R Markdown 100%

Node.js 80%

Jekyll 100%

Bootstrap 80%

Reproducibility & Containerization

Git 100%

Docker, Docker Compose 100%

CI/CD: GitHub Actions, cron 95%

Binder 100%

Jupyter Book 100%

Versioned data pipelines: DVC, DagsHub 60%

Cloud & Virtualization

Digital Ocean: droplet orchestration, backups, auto-deployment 90%

Google Cloud Platform: Compute, IAM, Storage 85%

Linux VMs: Ubuntu, WSL2, containerized apps 95%

Data Engineering & Storage

Non-relational databases: MongoDB 90%

Relational databases: SQL, MySQL 70%

REST APIs 100%

Deployment & MLOps

Infrastructure: JupyterHub, RStudio Server, multi-user platforms 100%

Web serving: NGINX reverse proxying, certbot, SSL 75%

Networking: dynamic DNS, WSL2 network bridging, access control 80%

Skills

My quantitative skills are built on seven years of advanced training in statistics and econometrics, spanning multiple quantitative Master's degrees and PhD qualifying exams in statistics (4.0 GPA). This foundation anchors my expertise in statistical modeling, computational methods, and rigorous empirical analysis.

Statistics

Regression analysis
Hypothesis testing
Time series forecasting
Model selection
Network analysis
ANOVA

Causal inference

A/B testing
Regression discontinuity design
Difference-in-difference
Instrumental variables
Panel data
Synthetic control
Interrupted time series
Matching and propensity score weighting

Machine learning

Classification: logistic regression, naive Bayes, support vector machines, K-nearest neighbor
Tree-based methods: decision tree, random forest, XGBoost
Feature engineering
Dimension reduction: principal components analysis, factor analysis
Regularization
Hyperparameter tuning
Model evaluation and loss metrics
Clustering: K-means, hierarchical, DBSCAN, spectral
Deep learning: artificial neural networks, generative adversarial networks, natural language processing
Causal learning

Computational methods

Constrained optimization
Expectation–maximization algorithm
Genetic algorithm
Sampling algorithms
Inter- and intrapolation
Numerical integration

Resume

For further detail on the skills, tools, and responsibilities involved in these roles, please consult my long-form CV.
Scroll down to download my one-page resume or my CV.

Education

PhD Environmental Science & Management

2016 - 2021

Bren School, University of California Santa Barbara, CA

Grade: 4.0. Qualifying exams: statistics and comparative politics.

MA Statistics

2017 - 2020

University of California Santa Barbara, CA

Grade: 4.0.

MA Economic History

2010 - 2011

London School of Economics, London, UK

Obtained with Merit (equivalent to magna cum laude in the US).

MSc Economics & Public Policy

2009 - 2010

Sciences Po Paris & École Polytechnique, France

Joint degree from France's top-ranked schools in social sciences and STEM.

BA (Hons) European Social & Political Studies

2005 - 2009

University College London, UK

Obtained with First Class Honors (equivalent to summa cum laude in the US).

Professional Experience

Executive Director

2024 - Present

Green Finance AI

Architecting and deploying end-to-end ML systems and infrastructure in cloud environments.
Providing technical consulting on machine learning and AI for external clients.
Delivering workshops and creating training content on generative AI and statistics.

Lead Machine Learning Engineer

03/23 - 06/23

Omdena (Global AI Collaboration)

Selected as part of a global collaboration of 20 engineers using AI for good.
Used LangChain and Chroma to develop context-aware chatbot for farmers seeking to optimize nitrogen applications.

Postdoctoral Fellow

2021 - 2023

Brown University, Watson Institute for International and Public Affairs

Worked on how to catalyze private sector investment in the green transition.
Built reproducible, containerized ML pipelines to support analytical research and model delivery.
Managed two Research Assistants and one developer.

Doctoral Researcher

2016 - 2021

Bren School of Environmental Science and Management, University of California Santa Barbara, CA

Devising innovative quantitative methods from econometrics and statistics to solve real-world, "wicked" problems.
Thesis: "A methodological toolkit to understand complex policy problems: applications to climate change and illicit finance".

Lead Instructor / Teaching Assistant

6 classes throughout 2017 - 2020

University of California Santa Barbara, CA

Classes include "Introduction to Research Methods" (undergraduate) and "Business & the Environment" (graduate).
Used innovative teaching strategies to introduce students to statistics, computer programming, and other technical concepts.
Consistently received rave reviews.

Principal Consultant

2015 - present

Various clients, including United Nations

Originated novel quantitative methodologies to trace illicit financial flows and uncover trade mis-invoicing across borders, combining investigative statistics with economic modeling.
Built global databases and user-facing tools to quantify country-level vulnerabilities and support global efforts on financial transparency.
Findings adopted by the African Union and used to inform UN policy efforts combating illicit trade in six countries.

Research Associate

2014 - 2015

Center for Global Development in Europe, London

Creator and lead developer of SkyShares, a simulation tool modeling the economic and environmental impacts of international climate deals, featured in The Guardian.
Designed algorithmic modules for carbon budget allocation, emissions trading, and cost-optimization; built full-stack app with Node.js, MongoDB, and interactive JS UI; released open-source code, methodology, and data for 190+ countries.
Directed cross-functional team of three; led policy outreach with UK and multilateral orgs.

Research Assistant

2012 - 2014

Center for Global Development in Europe, London

Helped launch European branch of leading DC-based economics think tank; wore multiple hats from research to IT setup.
Quantitative research on climate change and illicit finance.