Hi, I'm Sacha

Relational Database

• - Multiple work projects @ Enedis
  - Multiple academic projects (e.g, Landlord app)

• Built highly normalized database schemas from scratch, from Conceptual Data Models (CDM) to Logical (LDM) and Physical (PDM) implementations in SQL.

  Developed stored procedures, triggers, and automation jobs (VACUUM, backups, log purge…).

• Technologies: PostgreSQL, Oracle DBMS, PL/SQL, Python, Shell.

NoSQL Databases

• - 1 academic project

• Explored and learned to use NoSQL databases with MongoDB. Queried, inserted, and manipulated documents using MongoDB’s query language.

• Technologies: MongoDB.

Python ETL & Medallion Architecture

• - 1 Work Project @ Enedis

• Built and enhanced Python ETL pipelines to ingest and transform data from multiple sources (REST APIs, flat files, exports) into a Bronze / Silver / Gold medallion-style data warehouse

• Technologies: Python, SQL, Scheduled jobs.

Dataiku workflow

• - 1 Work Project @ Enedis

• Designed and maintained Dataiku workflows & recipes to automate ingestion, cleaning, preparation, and transformation of heterogeneous data sources.

• Technologies: Dataiku, Python, SQL.

KNIME Analytics Platform

• - 1 Academic project: OpenFoodFact data analysis

• Built an ETL pipeline using KNIME and Python to convert large NoSQL-like CSV datasets into structured and cleaned relational tables, for PostgreSQL

• Technologies: Knime, Python, SQL.

Power BI

• - Multiple work projects @ Enedis
  - 1 Academic project: OpenFoodFact data analysis

• Created interactive dashboards with automated refresh, data modeling, DAX measures, and workspace configuration.

  Worked with PBIP project structure, Git versioning, Azure DevOps and Power BI Service for production-ready analytics pipelines.

• Technologies: PowerBI, PBI Service, Git, Azure.

Oracle Application Express

• - 2 academic project

• Developed dashboards and interactive SQL reports to analyze student satisfaction and behavioral insights (e.g., Moodle connection and academic performance).

  Built dynamic charts, data summaries, useful filters and AI-assisted insights inside APEX.

• Technologies: Oracle APEX.

Revu — Incident cause classification

• - 1 Work Project @ Enedis (in progress)

• Predict the cause of an incident from a free-form comment (10–20 known categories).

  Dataset: >7000 comments · labeling target: ~500.

  Model: Transformers based embeddings + MLP classifier. Evaluation: confusion matrices + K-fold cross-validation. Delivery: Python API (batch + single prediction).

• Technologies: Pytorch, panda, MLP, Embedding, Fine tuning.

"Nutri-Score" prediction & visualisation

• - 1 Academic project: OpenFoodFact data analysis

• Built an end-to-end pipeline on OpenFoodFacts to clean and analyze data, then trained supervised models to predict nutritional scores, find out clusters and produce visual insights.

  Combined feature engineering, model evaluation, and dashboarding to explore correlations and explain predictions.

• Technologies: Python, panda, KNIME, PostgreSQL, Power BI.

Microscopic image analysis (grain size/shape/count)

• - 1 Academic project: Microscopic image analysis

• Implemented an image-processing pipeline (segmentation + export) to analyze metallic microstructures.

  Extracted per-grain features: area, perimeter, shape approximation (triangle/rectangle/square/oblong...), and size class (small/medium/large relative to the global average).

• Technologies: Python, NumPy, OpenCV, matplotlib.

OpenFoodFacts — LLM-based extraction

• - 1 Academic project: OpenFoodFact data analysis

• Used an LLM to extract nutritional data, excipients, and estimate the number of additives from picture of product packaging.

• Technologies: OpenAI API, Python.

Oracle APEX — AI summaries in dashboards

• - 1 Academic projects (Oracle APEX dashboards)

• Generated concise summaries of open-ended survey feedback using OpenAI API to complement interactive reports and charts directly inside APEX.

• Technologies: Oracle APEX, SQL, OpenAI API.

Local LLM inference (experiments)

• - Personal experimentation (local LLMs)

• Explored open models (8B–20B) for local inference, explored CPU/GPU offload trade-offs, and tracked throughput depending on setup.

• Technologies: Ollama, LM Studio, GGUF, quantization.

Python

• - Multiple work project @ Enedis
  - 2 academic projects: NutriScore prediction, Microscopic image analysis

• Used Python for data analysis, data processing, ML prototyping, and backend processing in academic projects. Also used for AI and scripting at work projects

• Technologies: Python, NumPy, pandas, scikit-learn.

Java

• - 2 academic projects: Landlord manager, Cheese seller

• Developed Java desktop applications with object-oriented design, MVC architecture, stock management and GUI (Swing / WindowBuilder).

• Technologies: Java, Swing, JDBC, Design Pattern.

C Language

• - 2 academic project: Text encryption

• Implemented encryption algorithms (Caesar, Vigenere) and built a command-line interface.

  Developed a C program for reversing list of password hashes, using rainbow tables attacks and lookup chains.

• Technologies: pointers, memory management, CLI, I/O.

PHP

• - 1 work project @ Enedis
  - Multiple academic projects (Handisport Manager S3/S4, )

• Built authentication systems, REST API backends, and complete CRUD applications in PHP.

• Technologies: PHP, MVC, REST, MySQL.

UI

• - 1 Work project @ Enedis
  - Multiple academic projects

• Built responsive UIs, optimized accessibility, and styled applications with modern CSS.

• Technologies: HTML, CSS, Javascript, Responsive design.

Dynamic content

• - 1 Work project @ Enedis
  - Multiple academic projects

• Developed dynamic interfaces (auth UI, dashboards, forms) for the Enedis webapp and my academic php projects.

• Technologies: Javascript, PHP templates, DOM, ajax.