Now — Data Analytics Engineer at L'Oréal Beauty Tech

Alexandre BREDILLOT

I turn complex data into I turn complex data into products people use, pipelines teams can trust, models that explain themselves and decisions backed by evidence.

Data & AI professional across Beauty Tech and Finance — from master data and data contracts at L'Oréal to payment data flows at La Banque Postale and FinBERT-powered return forecasting in my MSc thesis.

Scroll

01About

From raw data to products, pipelines and decisions.

Professional Summary

Data & AI Professional with a proven track record in building scalable data products across the Beauty Tech and Financial sectors. Experienced in engineering robust data pipelines, enforcing strict data quality standards, and driving statistical analysis. Skilled in Python, SQL, and BI tools, with a strong passion for transforming complex data into actionable business insights.

Based in Paris, France — mobile and open to relocation

Currently

L'Oréal Beauty Tech

Data Analytics Engineer in the Global Commerce Data Domain — built “Find Your Customer”, a React app on GCP, and formalising data contracts for priority Commerce data flows.

10M

customers searchable in “Find Your Customer”, the web app I designed — 100+ daily active users

+5.18%

MAE improvement at the 5-day horizon, statistically significant in my MSc thesis

21

certifications — IBM, Anthropic, Bloomberg & SAFe

5

languages — French, English, Spanish, Russian, Romanian

Beyond work

Travel blog

Stories, photos and maps — 1 destination on 1 continent so far, updated monthly.

  • Finance & investing
  • Calisthenics
  • Running
  • Motorsports

02Experience

Professional Experience

  1. July – December 2026 · Clichy, France

    L'Oréal — Beauty Tech Current

    Data Analytics Engineer – Global Commerce Data Domain

    • Data Product Build (React & GCP): Designed and deployed "Find Your Customer", a user-centered React web app on GCP (versioned on GitHub) that lets business and data teams search for a customer, view key information and easily access the underlying data from many sources that were previously scattered and unstructured.
    • Data Contracts & SLAs: Contributing to the implementation or formalization of data contracts for the priority data flows of the Commerce scope, defining quality rules, alerting and expected service levels.
    • Master Data Harmonization: Structuring massive volumes of complex Master Data (10M customers, 40k+ points of sale, 10k sales reps in the Sales Force Structure) from the SAP ECC and TIBCO EBX source systems into scalable, business-ready assets.
    • Data Quality & Remediation: Identifying and prioritizing personal data quality gaps on the Commerce scope, then steering the corrective actions with the IT teams.
    • Run & Operational Continuity: Handling requests and incidents on the assigned scope within agreed timeframes, keeping transparent follow-up on open topics and maintaining up-to-date documentation for recurring processes.
    • Cross-Functional Collaboration: Working daily with Data Product Owners and Data Engineers on the Global Commerce Transformation strategy, building a structured understanding of Commerce and Data dynamics to become an identified point of contact and share an actionable synthesis with the SDDS Commerce team.
    • GCP
    • GitHub
    • SAP ECC
    • TIBCO EBX
    • SQL
    • AI
    • Product Ownership
    • React
    • CI/CD
    • Python
    • Guardrails

    Manager: Mr. Luc Bochet, Lead Data Product Owner

  2. January – July 2025 · Paris, France

    La Banque Postale — Data Team, Payment Department

    Data Analyst

    • Data flow governance: supervised and optimized payment flow processes with focus on SEPA/DSP2/IP compliance.
    • Business needs analysis & implementation: partnered with stakeholders to understand needs, analyze processes, and deliver tailored reporting automation and KPI tracking improvements.
    • Automation & reporting: built dashboards and monitoring tools (QlikSense, Dataiku, SQL, Python) and automated recurring reporting to reduce manual workload and improve real-time visibility.
    • XML file mapping: mapped and structured XML files to improve traceability, usability, and downstream integration.
    • Data flow optimization: identified anomalies, improved processes, and supported operational reliability for payment workflows.
    • Data quality management: implemented validation rules and automated error-detection mechanisms to increase accuracy and consistency of financial data.
    • Process improvement: identified bottlenecks in data workflows and delivered optimizations improving processing time and operational efficiency.
    • QlikSense
    • Dataiku
    • SQL
    • Python
    • XML
    • SEPA / DSP2

    Manager: Mr. Hervé Harvoire, Payment Hub Manager

  3. January – July 2024 · Luxembourg

    Sienna Investment Managers — Finance Team

    Accounts Payable

    • Treasury management & reporting: executed daily treasury tasks, reconciled bank statements, and produced liquidity/cash reports; contributed to cash projections and month/year-end processes.
    • Management reporting & analysis: supported reporting and performed financial analysis to help monitor portfolios and operational performance.
    • Accounting & invoicing: processed and verified supplier invoices; managed sales invoices and maintained accounts payable/receivable; resolved discrepancies with vendors.
    • Regulatory compliance: supported VAT and Income Tax preparation and ensured adherence to internal policies and accounting standards.
    • Documentation & payments: maintained audit-ready records and supported payment processing (checks, ACH, wire transfers); reconciled employee expense reports.
    • Treasury
    • Financial reporting
    • Accounts payable
    • VAT

    Manager: Ms. Aija Livca, Account Payable Manager

  4. April – July 2023 · Issy-les-Moulineaux, France

    EasyBourse — Business Development Team

    Business Development Assistant

    • Conducted market and sales performance analysis to identify trends and opportunities.
    • Coordinated incentive programs, email campaigns, and digital advertising to support client engagement.
    • Supported commercial activity by preparing quotations and assisting with business offers.
    • Market analysis
    • Email campaigns
    • Digital advertising
  5. Since December 2022 · Fontevraud L'Abbaye, France

    French Army Reserve — 2nd Dragoon Regiment, 6th Squadron Ongoing

    Reserve Soldier

    • Reserve soldier in the 6th squadron of a cavalry regiment specialized in NRBC.
    • Trained and equipped to operate in and respond to Nuclear, Radiological, Biological, and Chemical (NRBC) threat environments.
    • Experience in high-discipline operational settings, teamwork, and readiness for time-constrained missions.
    • NRBC

03Projects

Selected projects

Featured · MSc Thesis · EDHEC 2026

Sentiment Analysis & Short-Term Return Predictability in Small-Cap Nuclear Equities

Does FinBERT-derived news sentiment improve short-term return prediction for small-cap nuclear equities? This MSc thesis builds a fully reproducible five-phase pipeline — feasibility audit, data engineering (3,992 ticker-days, 25,885 news articles covering SMR, LEU, NNE), FinBERT sentiment scoring, LSTM training with chronological hold-out, and SHAP explainability. Key finding: sentiment augmentation yields a statistically significant average MAE improvement of +5.18% at the 5-day horizon; no aggregate benefit at the 1-day horizon. Hypothesis testing via Diebold-Mariano tests.

  • Python
  • FinBERT / NLP
  • LSTM
  • SHAP
  • Diebold-Mariano
MAE gain at 5 days
+5.18%
news articles
25,885
ticker-days
3,992
tickers · SMR, LEU, NNE
3

Five-phase reproducible pipeline

  1. 01Feasibility audit
  2. 02Data engineering
  3. 03FinBERT sentiment scoring
  4. 04LSTM · chronological hold-out
  5. 05SHAP explainability

Folio — Nuclear Portfolio Dashboard

A personal Streamlit dashboard for a nuclear-heavy equity portfolio, built to check from a phone or laptop. It values positions in EUR (splitting stock moves from currency moves), scores the broad market, the nuclear sector and the portfolio on a −100 to +100 bearish-to-bullish scale (trend, momentum, RSI, CNN Fear & Greed, VIX), and sets bull, base and bear 12-month scenarios with volatility cones and vol-implied odds. News tone (VADER), StockTwits and Reddit hype round it out; the app keeps working when a data source is down, and the calculation code is unit-tested.

  • Python
  • Streamlit
  • Plotly
  • yfinance
  • VADER

TRAD_BOT — Algorithmic Trading Suite (Binance + IBKR)

A Python trading suite covering Binance funding-rate arbitrage, a BTC trend bot with a Streamlit dashboard, and IBKR equity sentiment trading. The latest addition is a small-cap crypto sentiment research bot: it screens Binance spot pairs using CryptoPanic/RSS news and a two-stage FinBERT → LLM pipeline, with spread and conviction filters, position limits, stop-losses and a daily-loss stop. The new bot is long/flat, spot-only and defaults to paper trading; its trading edge remains unproven.

  • Python
  • Binance / IBKR
  • asyncio
  • SQLAlchemy
  • Streamlit

n3gh — Open, Science-Based Health Data

An open-source project making health data available to people looking for evidence-based advice and recommendations, instead of ad-driven content. Ships as a small, fast static site (GitHub Pages) with a comparison tool as its first building block, plus a bot module handling data automation behind the scenes. Actively evolving — more science-based tools are on the way.

  • HTML/CSS/JS
  • Open Source
  • GitHub Pages

dailynews — A Personal Newspaper for Kindle

A personal daily newspaper built from my interests and delivered to my Kindle every morning, running free on a GitHub Actions schedule — no server, no app. Each run collects recent articles from a news search per interest plus RSS feeds, drops anything already sent, deduplicates the same story across outlets, ranks the best pieces per section, extracts the full text (respecting robots.txt; paywalled pieces fall back to summary + link), builds a clean EPUB with a clickable front page and emails it via Send to Kindle. Tested with pytest and linted with ruff in CI.

  • Python
  • GitHub Actions
  • feedparser
  • trafilatura
  • EPUB

More in the pipeline

This space is reserved for upcoming work in data science, quantitative finance and AI. Stay tuned — new projects are on the way.

View Projects on GitHub

04Education

Education

Since 2025 — Lille, France

EDHEC Business School

MSc Data Science & Artificial Intelligence

Ranked #4 Business School in France. Focus on applied data science and AI for business.

  • Core: Python for data analysis; statistical modeling & inference; data mining & machine learning (clustering/PCA, classification, ensembles), model validation & interpretability.
  • Advanced: Deep learning (Keras), NLP (topic modeling, embeddings, sentiment + intro to LLMs), time series forecasting (ARIMA/SARIMA + ML), scalable data management & analytics engineering (SQL/cloud pipelines, production considerations).
  • Applied analytics: Web analytics stack (GA4/GTM, BigQuery SQL, Looker Studio) and business projects translating data into decision-ready dashboards and recommendations.

2022 — 2025 — Paris, France

ESSCA School of Management

Bachelor in International Management

Studied international management with coursework in finance and analytics.

  • Business, strategy & project delivery: project management, strategy, organisational behaviour, cross-cultural management, business simulation.
  • Marketing, negotiation & communication: marketing project, negotiation, market research, multimedia communication (campaigns & stakeholder comms).
  • Finance, economics & governance: finance for managers, economics, political economy, international trade, intro to law & international relations.

05Skills

Skills

Programming & Data

  • Python
  • pandas
  • NumPy
  • scikit-learn
  • PyTorch
  • Keras
  • matplotlib
  • seaborn
  • SQL
  • PostgreSQL
  • VBA
  • HTML / CSS
  • React
  • Git / GitHub

Analytics & BI Platforms

  • Dataiku
  • QlikSense
  • Tableau
  • Power BI
  • Looker Studio
  • BigQuery
  • Google Cloud (GCP)
  • GA4 / GTM
  • Excel (financial modelling)

Methods & Soft Skills

  • Data governance
  • Data quality
  • Data contracts
  • Master Data Management
  • Personal data compliance
  • Run / incident management
  • Statistical analysis
  • Time-series forecasting
  • NLP / Sentiment analysis
  • Model interpretability (SHAP)
  • Product thinking
  • Cross-functional collaboration
  • Mentoring

Languages

  • FrenchNative
  • EnglishAdvanced (C1)
    TOEIC 970 · TOEFL iBT 95
  • SpanishIntermediate (B2)
  • RussianElementary (A2)
  • RomanianBeginner (A1)

Certifications

Anthropic8
  • AI Capabilities and Limitations
  • Model Context Protocol: Advanced Topics
  • Claude Code in Action
  • Introduction to Claude Cowork
  • Claude Code 101
  • Claude 101
  • Building with the Claude API
  • Introduction to Model Context Protocol
IBM11
  • What is Data Science?
  • Tools for Data Science
  • Python for Data Science, AI & Development
  • Machine Learning with Python
  • Data Science Methodology
  • Python Project for Data Science
  • Databases and SQL for Data Science with Python
  • Data Scientist Career Guide and Interview Preparation
  • Data Visualization with Python
  • Data Analysis with Python
  • Generative AI: Elevate Your Data Science Career
SAFe by Scaled Agile Certified SAFe® 6 Agilist 2026
Bloomberg Bloomberg Market Concepts 2025

Interests

  • Finance & investing
  • Calisthenics
  • Running
  • Travelling
  • Motorsports
  • Video games

06Contact

Let's work together

Open to opportunities in data science, AI strategy, and data governance roles. Feel free to reach out!

Send me a message — I usually reply within a day.