Bienvenue sur le profil Malt de Pierre !

Malt vous permet d'accéder aux meilleurs consultants freelances pour vos projets. Contactez Pierre pour échanger sur votre besoin ou recherchez d'autres profils sur Malt.

Pierre Canavelli

machine learning scientist & data engineer

En télétravail depuis Paris

  • 48.8546
  • 2.3477
Proposer un projet La mission ne démarrera que si vous acceptez le devis de Pierre.
Proposer un projet La mission ne démarrera que si vous acceptez le devis de Pierre.

Localisation et déplacement

Localisation
Paris, France
Télétravail
Effectue ses missions majoritairement à distance

Préférences

Secteur d'activité
Préfèrerait éviter:
  • Grande distribution
  • Presse & médias
  • Ressources humaines
Taille d'entreprise
Préfèrerait éviter:
  • 1 personne
  • 2 - 10 personnes

Vérifications

Langues

Catégories

Compétences (25)

  • Débutant Intermédiaire Confirmé
  • Débutant Intermédiaire Confirmé
  • NLP
    Débutant Intermédiaire Confirmé
  • Frameworks
  • Débutant Intermédiaire Confirmé

Pierre en quelques mots

I a PhD machine learning scientist and graduate from the Ecole Normale Supérieure of Paris with 7 years' of experience applying AI to academic research and real-world problems. I am used to working with difficult data (e.g. terabyte-sized, extremely imbalanced or dirty datasets) spanning a variety of industries, including financial timeseries; highly sensitive, anonymized health records; and noisy, low-resolution video. I have experience working with tasks covering fraud detection, predictive maintenance, rare disease diagnosis, video processing, and customer sentiment analysis.

I do most of my ML work in Python, SQL, and Julia, and implement models using TensorFlow, SKLearn, and a handful of common libraries such as XGBoost, LightGBM, and CatBoost. I am adamant about following software engineering best practices: all production code must be modular, unit-tested, documented, human-readable, and adhere to the Google coding standards.

Having spent most of my career doubling as a data engineer, I am also well-versed in setting up, expanding, and maintaining data mining, ETL, feature engineering, and warehousing pipelines using PySpark and Dask frameworks. I also have a strong experience in working with real-time data streaming pipelines using Kafka, Faust and ClickHouse.

Due to my past experience working as a technical consultant for major management consulting firms, I am a clear and concise communicator, and welcome client-facing activities. I particularly enjoy working in diverse, small-to-medium-sized teams alongside colleagues hailing from different backgrounds and fields of expertise.

Expériences

Shadow

High tech

Machine Learning Scientist & Engineer

Paris

janvier 2020 - Aujourd'hui (1 an et 10 mois)

Shadow is building the future of personal computers. I joined them with a double mission: to expand their data engineering stack, and to bring machine learning to the company. Current activities include:

- Expanding, improving, and maintaining our data streaming pipelines using Kafka, Faust and ClickHouse.

- Democratising access to data by creating interactive dashboards and accessible endpoints using Metabase and FastAPI.

- Leading AI/ML projects spanning anomaly detection, NLP, causal analysis and supervised learning.

- Providing expertise in deep learning and engineering best practices on R&D projects focused on real-time video data.

- Mentoring analysts, data scientists and R&D engineers.

- Writing and maintaining Spellbook, our in-house ETL and ML/DL library.

IQVIA

Industrie pharmaceutique

Data Scientist

London, United Kingdom

mai 2019 - octobre 2019 (5 mois)

IQVIA is the world leader in healthcare information technologies. I had the pleasure to join their Predictive Analytics team, where I worked on developing new machine learning models for the early diagnosis of rare diseases. It was all going great. But then, Brexit happened.

- Designed and delivered machine learning models for the predictive diagnosis of rare diseases among the US population.

- Wrote and maintained Python packages automating the ETL, feature engineering and QC of terabyte-scale datasets.

- Provided ad hoc statistical data analysis for US- and UK-based healthcare consulting teams.

Resonate

Transports

Data Scientist

London, United Kingdom

novembre 2018 - mai 2019 (6 mois)

Resonate's mission is to improve the efficiency, safety and sustainability of the rail industry using data science and machine learning. I worked with Resonate's Data team as a full-stack data scientist, splitting my time between designing, training and deploying predictive algorithms, and building up our data engineering stack.

- Trained and deployed machine learning models for the forecasting and prevention of disruptive events on the National Rail network.

- Designed and implemented end-to-end machine learning pipelines allowing for the automated data preparation, training and evaluation of ML models on EC2 instances.

- Initiated a 4-people effort to refactor, optimise and document 5,000+ lines of legacy code, leading to performance improvements up to 200x and full compliance with Google coding standards.

- Implemented and maintained data ingestion pipelines using AWS S3, Glue, Athena, QuickSight, and SPICE, delivering customer-facing dashboards and static reports.

Faculty

Data Science Fellow

London, United Kingdom

septembre 2018 - novembre 2018 (2 mois)

1 recommandation externe

Consultez les recommandations qu'a reçues Pierre

Formations