Clément Sicard
🇫🇷 data professional living in 🇨🇭, with international experience in {🇺🇸,🇨🇭,🇫🇷,🇸🇪}
About
With a background in theoretical machine learning and experience in Big Data and software engineering, I strive in both AI/ML and Data Engineering roles. I have worked for 5+ years on a wide range of technical projects in a variety of different fields: giga-scale battery manufacturing, academic research, consulting, environmental services, financial services, startups and international organisations. I aim to leverage my experience to help companies build and deploy data products, while having a positive impact on society, especially on social justice and climate change mitigation. My research interests focus on deep learning and its applications to multimodal data for action recognition tasks, with a stress on audio.
Work Experience
Isomorphic LabsLausanne, 🇨🇭London, 🇬🇧
Data Engineer
New York UniversityNew York, 🇺🇸
Deep Learning Research @ Music & Audio Research Lab (MARL)
United NationsNew York, 🇺🇸Remote
NLP Research
NorthvoltStockholm, 🇸🇪Remote
Data Science/Engineering Consultant
swissQuant GroupZurich, 🇨🇭
Quant Engineer (Software & Maintenance)
Green PRAXISMarseille, 🇫🇷Remote
Lead Engineer (Software & Data)
Ecole Polytechnique Fédérale de Lausanne (EPFL)Lausanne, 🇨🇭
Research Assistant @ Digital & Cognitive Musicology Lab (DCML)
Education
ETH Zurich
New York University
Ecole Polytechnique Fédérale de Lausanne
Skills
Projects
SemUN
A full-stack semantics-powered search platform for the United Nations' Digital Library
Spaiche: Extending State-of-the-Art ASR Models to Swiss German Dialects
State-of-the-art model for Swiss German dialects speech recognition, with a novel loss function based on semantic distance in the embedding space. Presented at SwissText 2023
USwinBaseNet
Combining visual transformers and U-shaped networks for segmentation tasks
opti-stem-programs
Machine-learning-assisted creation of optimally located STEM programs in the city of Chicago
undl
A lightweight Python library to inreact with the United Nations' Digital Library API
air-quality-forecasting
Using data we measured using sensors linked to a Raspberry Pi, we derived an LSTM model to predict air quality in Zurich and compared its performance with other more basic models
un-ml-pipeline
Machine Learning pipeline and API for the United Nations Digital Library
fgsm-attacks
Targeted and untargeted FGSM attacks on pre-trained Neural Networks
pgd-trades-defenses
PGD and TRADES defenses for adversarial-attacks-resistant Neural Networks
GameBoj
End-to-end GameBoy emulator developed from scratch in Java, from hardware specifications to software implementation
tetris-nios
Tetris implementation on Gecko FPGA boards, using NIOS II softcore processor
AutoMove
Python script running in background for macOS, to automatically move Notability backups from either a OneDrive directory or a NAS to an iCloud directory, in realtime.
constrained-dtw
Python implementation of constrained Dynamic Warping Algorithm
StudyHelper
StudyHelper is a mobile (iOS, Android) and desktop (Windows, macOS, Linux) application that aims at helping learning course material with the mean of flashcards
WhereIsWaldo
Java program able to locate Waldo on given images, using pattern recognition and basic image processing
PyACME
Implementation of an ACMEv2 (RFC-8555) client in Python, to automate the process of issuing and renewing TLS X.509 certificates with Let's Encrypt.
SeniorLauncher
An easy-to-use and lightweight Android launcher designed using Flutter, for seniors who are not familiar with technology
c-priority-queues
Generic priority queues implementation in C language.