§ About · curriculum vitæ
Curriculum vitae
§ 01
Experience
Oct 2024 - Present
Data and Digitalisation Senior Scientist
Roche · Basel, Switzerland
- Lead of the Data, Digitalization & AI Cross-Functional Circle (D2AI) at pCMC in Small Molecule Research
- Developed APPA (Agentic Preformulation Pathway Assistant), an LLM-based agentic system that automates multi-step drug preformulation workflows
- Built ML reranking pipeline for crystal structure prediction, reducing polymorph search space by orders of magnitude (published CrystEngComm 2024)
- Driving adoption of ML potentials and descriptor-based models across formulation and solid-state chemistry teams
Mar 2023 - Oct 2024
Data and Digitalisation Scientist
Roche · Basel, Switzerland
- Benchmarked molecular descriptors (SOAP, MBTR, Coulomb matrices) for solubility and lipophilicity prediction across pharmaceutical compound libraries
- Developed solute-lipid interaction models connecting atomistic descriptors to formulation performance
- Built interactive web dashboards (React, Next.js) for real-time visualisation of molecular property landscapes
- Co-authored publications in Molecular Pharmaceutics and CrystEngComm
Jan 2021 - Mar 2023
Roche Postdoctoral Fellow
Roche · Basel, Switzerland
- Applied high-dimensional ML potentials and committee-of-experts models to pharmaceutical crystal polymorph screening
- Developed Bayesian NMR crystallography methods for solid-form identification under uncertainty
- Published work on ML-accelerated crystal structure prediction and molecular crystal energy ranking
- Bridged academic ML method development with industrial drug development pipelines
Oct 2020 - Jan 2021
Economic Analyst
Pictet Group · Geneva, Switzerland
- Built automated data pipelines for Pictet Asset Management's database infrastructure upgrade
- Applied quantitative analysis methods to financial time-series data
May 2016 - Jul 2020
Doctoral Assistant — PhD in Materials Science
EPFL (École polytechnique fédérale de Lausanne) · Lausanne, Switzerland
- Designed divide-and-conquer committee-of-experts architecture for high-dimensional machine learning potentials (published Digital Discovery 2024)
- Developed automatic feature selection methods for atomic fingerprints (SOAP, ACSF) with extrapolation diagnostics
- Built Python/C++ frameworks for training and deploying neural-network potentials on condensed-matter systems
- Contributed to the open-source librascal library for atomistic ML representations
- Thesis: ML architectures for characterisation and classification of molecular materials
Mar 2015 - Oct 2015
Visiting Research Graduate Student
Massachusetts Institute of Technology (MIT) · Cambridge, MA
- Developed molecular dynamics simulations of ethanol nanofiltration through nanoporous graphene membranes
- Built computational models for selective molecular transport in 2D materials
- Collaborated with Prof. Grossman's group on computational nanomaterials research
Sep 2014 - Jan 2015
Special Research Student
Tohoku University · Sendai, Japan
- Simulated carrier mobility variations in vertical MOSFET architectures using TCAD tools
- Optimised semiconductor device geometries through systematic computational parameter sweeps
§ 02
Education
2016 - 2020
Doctor of Philosophy — PhD, Materials Science
EPFL (École polytechnique fédérale de Lausanne) · Lausanne, Switzerland
- Thesis: Machine learning architectures for characterisation and classification of molecular materials
- Developed committee-of-experts potentials and automatic descriptor selection for atomistic ML
- Published in Digital Discovery, Physical Chemistry Chemical Physics, Acta Crystallographica
- Advisor: Prof. Michele Ceriotti (Laboratory of Computational Science and Modelling)
2013 - 2015
Master of Science — MS, Nanotechnology applied to ICTs
Politecnico di Torino · Turin, Italy
- Final grade: 110/110 cum laude
- Research stays at MIT (7 months) and Tohoku University (5 months)
- Focus on computational modelling and simulation of nanoscale systems
§ 03
Capabilities
Machine Learning PotentialsCrystal Structure PredictionGraph Neural NetworksPyTorchPythonC++Molecular DynamicsAtomistic Representations (SOAP, ACSF)LLM AgentsBayesian MethodsActive LearningComputational ChemistryDrug Discovery & FormulationReact / Next.jsScientific Computing