Skip to content
View albertodiazdurana's full-sized avatar
๐ŸŽฏ
Focusing
๐ŸŽฏ
Focusing

Block or report albertodiazdurana

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.

Maximum 250 characters. Please don't include any personal information such as legal names or email addresses. Markdown supported. This note will be visible to only you.
Report abuse

Contact GitHub support about this userโ€™s behavior. Learn more about reporting abuse.

Report abuse
albertodiazdurana/README.md

Hi, I'm Alberto Diaz Durana

Senior Data Scientist | 10+ Years Experience | Berlin

Building production ML systems across manufacturing, sustainability, and financial services. Currently specializing in LLM integration and agentic AI at Masterschool.

Domain expertise: Energy Systems Engineering (DIN/VDI standards) | Process Mining & BPM Optimization

LinkedIn Portfolio


Featured Projects

Project Description Tech App
RAG Document Assistant (In Development) Production RAG with multi-provider LLMs and vector databases LangChain, ChromaDB, FastAPI, MLflow -
DevFlow Analyzer Agentic AI + process mining for CI/CD analysis (10K+ builds, 86 tests) LangChain, LangGraph, PM4Py, MLflow Live
Disaster Tweet Classification NLP evolution: TF-IDF โ†’ Embeddings โ†’ Transformers (F1: 0.77) Sentence Transformers, GloVe, FastText -
Computer Vision Manufacturing defect detection; U-Net, autoencoder, ResNet TensorFlow, MLflow, Streamlit Live
Residential Energy Systems Domain reference: ML + German heating standards DIN/VDI/GEG, MLOps, Time Series Live
Demand Forecasting 4.8M transactions, 11% RMSE improvement XGBoost, MLflow, TensorFlow, Streamlit Live
Customer Segmentation $23M CLV analysis, 97.2% confidence scikit-learn, pandas, PCA -

View Full Portfolio

Tech Stack

ML & AI: TensorFlow | XGBoost | LangChain | LangGraph | Sentence Transformers | scikit-learn | SHAP | MLflow

Data: Spark | Polars | pandas | SQL

Deploy: Streamlit | FastAPI | AWS


๐Ÿ’ผ Professional Highlights

  • Alcemy GmbH (2024-2025): Deployed 5+ ML models optimizing cement production, cutting COโ‚‚ emissions across 35+ customers
  • Appian Software (2021-2024): Led 10+ process mining assessments, reducing process times ~20% on average
  • TU Berlin (2019-2021): PhD research in energy access prediction; CPOTE 2020 publication
  • HEDERA Sustainable Solutions GmbH (2018-2021): Co-founded sustainability startup, built cloud-based systems for 15+ international projects

๐ŸŽ“ Education & Certifications

  • Masterschool โ€” AI & Data Science Program (08.2025-03.2026)
  • DeepLearning.AI โ€” MLOps Specialization (2024)
  • future Training & Consulting GmbH โ€” Data Science with Python (2018)
  • Data Analytics for Six Sigma (University of Amsterdam, 2017)
  • PMI โ€” Project Management Professional - PMP (2016)
  • TU Berlin โ€” MSc Process, Energy & Environmental Systems Engineering (2010-2013)
  • Universidad de los Andes โ€” Mechanical Engineering Diploma (2001-2006)

๐ŸŒ Languages

Spanish (Native) | English (C2) | German (C2) | Portuguese (B2)


Open to opportunities in Data Science, ML Engineering, and AI Product Development

Email LinkedIn

Pinned Loading

  1. data-science-portfolio data-science-portfolio Public

    Senior Data Scientist portfolio: time series forecasting, process mining, customer analytics, and industrial ML. Bridging mechanical engineering and AI with expertise in energy systems optimizationโ€ฆ

  2. CorporacionFavorita-demand-forecasting-in-retail CorporacionFavorita-demand-forecasting-in-retail Public

    End-to-end demand forecasting pipeline for retail inventory optimization; 4.8M transactions, XGBoost RMSE 6.40, MLflow tracking, Streamlit deployment

    Jupyter Notebook

  3. Demand-forecasting-in-retail-app Demand-forecasting-in-retail-app Public

    Interactive demand forecasting app for Corporaciรณn Favorita - Streamlit deployment

    Python

  4. loan-approval-prediction loan-approval-prediction Public

    Machine learning project predicting loan approval based on applicant and financial details.

    Jupyter Notebook