Adam R.

Compétences

• NLP5/5
• Python5/5
• LLM5/5
• Machine Learning5/5
• Vue.js2/5
• Scala3/5
• FastAPI3/5
• AWS2/5
• SQL4/5
• Pandas5/5
• Spark4/5
• TensorFlow3/5
• PyTorch5/5
• Elasticsearch3/5
• RAG5/5
• scikit-learn5/5
• Fine-tuning5/5
• Deep Learning5/5

Expériences

1. Phone

   Jan 2024 — Oct 2026

   • Sorbonne Université, Email • LIMICS, Paris, France ad.remaki@gmail.com

2. Location

   Nov 2021 — Jan 2024

   • Greater Paris University • Website • Hospitals (AP-HP) https://aremaki.github.io

3. London, UK

   Sep 2017 — Oct 2021

   • Frameworks: Scikit-learn, PyTorch, TensorFlow, Paris, France • Transformers, Pandas, Polars, PySpark, SpaCy, FastAPI, TRL.

   • FastAPI
   • Pandas
   • TensorFlow
   • PyTorch
   • scikit-learn

4. Data Science: Data Processing

   Jan 2026 — May 2026

   • Learning, Model Training & GitHub - HuggingFace • Evaluation, Hyperparameter • Paper accepted at EACL • Optimization, Fine-tuning.

   • Fine-tuning

5. Statistics, Data Analysis &

   Nov 2022 — Nov 2023

   • Machine Learning. • PhD Researcher — NLP & Data Science • Research in clinical NLP, machine learning and generative AI, with a focus on extracting and normalizing information from clinical notes for large-scale healthcare data reuse. • Data Scientist • Applied data science to large-scale Electronic • Health Record data from Europe's largest clinical data warehouse, developing open-source Python libraries to support data science on the platform. • Full-Stack Developer • Developed and maintained HelloFresh’s core operations platform, led its rollout in New Zealand, and built an AWS-hosted recipe planning system to optimise operational KPIs. • MSc in Applied Statistics • Imperial College London — Distinction • Engineering Degree • CentraleSupelec (formerly Centrale Paris) • LongBEL (entity linking) • Document-level biomedical entity linking using document context and prediction memory to improve consistency across clinical texts. • SynCABEL (entity linking) • Synthetic data generation with LLMs to improve entity linking and reduce annotation efforts. • BioMedics (Information Extraction) • NLP pipeline that transforms unstructured French clinical notes into structured data. It performs NER, qualification, entity linking and relation extraction. • EDS-TeVa (Statistical Analysis) • Python library to model data availability over time and across hospitals, helping identify and mitigate temporal bias in clinical data warehouse analyses. • Selected Projects

   • Python
   • AWS
   • Machine Learning
   • NLP
   • LLM