PIFreelancer profile translated to English.
Back to original languageAbout Pety Ialimita
Did you know?
85% of data-driven projects fail to create value, often due to a lack of understanding of business issues and dialogue between tech and decision-making teams, according to RAND corporation in 2024.
My background in a nutshell
A graduate of Télécom Paris and in Economics, with experience at Volvo, Idemia, Veolia, BRED, I combine cutting-edge expertise in data science/AI with a strategic business vision. This dual skill set allows me to frame the right problems and design tailor-made data solutions, adapted to your constraints and objectives.
Tailor-made approach:
Depending on the nature of the use case and the level of explainability required, I adapt the method:
- Statistics or econometrics for simple and interpretable models;
- Advanced AI (ML, deep learning, LLM) for processing massive volumes or real-time data.
The choice of method, or at least the project protocol from the start, between us, will maximize your return on investment.
You remain in control:
- Choice of technologies such as Python or R;
- Level of automation, such as offline trained models or online retraining;
- Scope of intervention (end-to-end modeling or targeted: exploration, training, retraining, deployment, etc.).
To optimize my impact, I focus on my areas of expertise:
- Forecasting & Nowcasting (time series, incremental learning)
- Risk/Anomaly Detection
- CRM & Customer Segmentation
- Computer Vision & Visual-based AI
- Sentiment or Document Analysis (NLP & LLM)
French
Native or bilingual
Remote only
Primarily works remotely
Experience
- Volvo Group FRData ScientistSeptember 2024 - February 2025 (5 months)Lyon, France
- Real-time anomaly detection during truck stress tests.
- Embedded AI: retraining after drifts, Forecasting
- Tools: Python / Jupyter notebook / Microsoft Azure
- IDEMIAArtificial Intelligence EngineerMarch 2024 - July 2024 (4 months)Paris, France
- Prediction of facial occlusion rate using a pre-trained CNN from Hugging Face.
- 1st place in the data challenge
- Solution: Ensemble Learning (ResNet-18) on color/BW images and weighting by occlusion level.
- Skills: Deep Learning Finetuning / Computer Vision / Dropout
- Télécom ParisData ScientistJanuary 2024 - March 2024 (2 months)
- Multi-class classification (28 classes) biased by gender.
- Resampling + bagging achieved 82.9% accuracy and reduced TPR_GAP (bias).
- Skills: Fairness · Scikit-learn · Ensemble Learning · BERT
Recommendations
Be the first to recommend Pety Ialimita
Help this freelancer shine by sharing your experience working together.
These freelancer profiles also match your criteria
Agatha Frydrych
Backend Java Software Engineer
4.7
(3)
2
Baptiste Duhen
Fullstack developer
4.6
(4)
5
Amed Hamou
Senior Lead Developer
4
(2)
7
Audrey Champion
Web developer
4.3
(3)
4
Education
- Advanced Master's Degree in Data & Artificial Intelligence ExpertTélécom Paris2025Atouts du Mastère Spécialisé : - Classé dans le top 03 des meilleurs formations en Big Data et Data Sciences en France depuis au moins 2023 ; - pédagogie progressive sur le traitement des données : Exploration Data Analysis -> Statistiques fondamentales -> Statistical Learning -> Data Challenge de classification multiclasse sur données biaisées ; -> Machine Learning avancé, Data challenge sur la prédiction du pourcentage de visage masqué (ou occulté) -> NLP et LLM, Projet de ZeroShot-Learning pour une banque ; - Projet final : -- Analyse de sentiment (positif/neutre/négatif) des bilans financiers et rapports économiques -- Prétraitement, vectorisation et inférence avec modèles Transformers. -- Compétences : Zero-shot learning, Onyxia (cloud), Pytorch, Hugging Face - Bagage théorique conséquent en Data science et adaptation aux nouvelles formes d'IA ; - Bagage solide en informatique : Python, SQL, Github, Cloud Computing, Système distribué
- Double Master's Degree II in Applied EconomicsUniversity of La Réunion / Catholic University of Madagascar2024- Major de promotion & Boursier d'excellence - Recherches : indépendance de la banque centrale & inflation - Compétences sur l'économétrie : Vector Autoregressif (VAR), Vector Error Correction Model (VECM)