Skip to content
View AchrafEssaleh's full-sized avatar

Block or report AchrafEssaleh

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Content in all repositories owned by your account will be closed.
Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
AchrafEssaleh/README.md

Achraf Essaleh

Final year engineering student (M.Eng) @IMT Atlantique · B.Sc. Mathematics & Computer Science @Sorbonne Université

Quantitative data science · Machine learning · Time series

Side projects: quantitative ML data challenges

Personal participation in public data challenges set by quantitative investment firms.

CFM challenge: End of Session Return Prediction · Ongoing · Python, scikit-learn

  • 3 class classification of US equity end of session return direction from 53 intraday five minute returns.
  • Statistical feature engineering and strict train/test separation across days and stocks to avoid leakage between market regimes.

QRT challenge: Market Neutral Alpha Prediction · Python, scikit-learn, LightGBM

  • Leakage aware pipeline on anonymised financial time series: missingness diagnosis, imputation, quality checks.
  • Volatility, momentum and reversal features from residual returns.
  • Benchmarked logistic regression, random forest and LightGBM; feature importance and ablation studies to validate the engineered signals.

Industry project

AI Assistant for Technical Knowledge Management, in collaboration with SOLENT

  • On premise RAG pipeline (ChromaDB, Jina v3) over Redmine data.
  • RAG Fusion retriever (multi query expansion, RRF) with Self RAG to filter out of context answers.
  • Local LLM via Ollama: no SaaS dependency, client data stays on premise.

Private repository.

Experience

Data Scientist Intern, (Portfolio Management team) · Paris

  • Built a constrained decision tree engine from scratch in Python, prioritising explainability.
  • Designed a modular OOP architecture with an interactive Plotly interface used by business analysts.
  • 82% test accuracy predicting client profitability on a 52K row production portfolio dataset.
  • Deployed to production via Databricks Apps.

Code is proprietary and not published.

Activities

  • Head of the Financial Engineering Division, KRYPTOSPHERE (2025 – Jan. 2026)

Coursework and other projects

Machine learning

  • MLOps for Clinical Trials (IMT Atlantique): end to end pipeline for drug approval prediction with MLflow, Airflow and Docker (experiment tracking, orchestration, reproducible deployment).
  • Spotify Top Hits, unsupervised learning (IMT Atlantique): PCA and K-Means on 2,300 tracks.
  • Sleep Quality Detector (personal): sleep pattern prediction from smartphone sensor data.

Systems, simulation and algorithms

  • AI Strategy for PyRat (IMT Atlantique, Python): heuristic and reactive agents for dynamic pathfinding.
  • Virtual Throwing & Haptic Feedback (IMT Atlantique): VR motion tracking and feedback loops.
  • Solar Irradiance Measurement (IMT Atlantique): Arduino-based real-time sensor system.
  • Jewel Collection Simulation (Sorbonne, Java): multi-agent OOP simulation.
  • Predator Simulation (Sorbonne, C): ecosystem simulation on a grid.
  • Automata Operations (Sorbonne, Python): finite automata construction and determinization.

Contact

LinkedIn · achraf.essaleh@imt-atlantique.net

Pinned Loading

  1. Market-Neutral-Alpha-Prediction Market-Neutral-Alpha-Prediction Public

    Jupyter Notebook 1

  2. mlops-project-main mlops-project-main Public

    MLOps project integrating Airflow, Docker, and MLflow for machine learning workflow automation

    Jupyter Notebook

  3. Spotify-sHits-Unsupervised_Learning_Challenge Spotify-sHits-Unsupervised_Learning_Challenge Public

    Forked from Elakiwala/Challenge-Spotify

    Spotify Trend Charts Analysis

    Jupyter Notebook 1

  4. end-of-session-return-prediction end-of-session-return-prediction Public

    Jupyter Notebook