~/anass

I build production AI systems and multi-agent architectures. Today that's at MediaMarktSaturn, automating retail and service workflows across international markets. Before that, Amazon Science & Tech.

View projectsResume

Experience

  1. AI Engineer

    02/2026 – Present

    MediaMarktSaturnNow

    Production AI systems and multi-agent architectures for retail and service operations, automating business workflows across international markets.

    • n8n
    • LangChain
    • Google ADK
    • CX Agent Studio

  2. SDE Intern, Science & Tech

    03/2025 – 08/2025

    Amazon

    Production RAG pipelines, multi-agent systems and ETL workflows for code understanding and knowledge retrieval across internal codebases, wikis and proprietary ML frameworks, cutting onboarding time and support overhead for thousands of internal users. Led the prototyping and benchmarking of agentic LLM assistants to improve reasoning and reliability while keeping hallucinations low.

    • RAG
    • Multi-agent systems
    • ETL
    • LLM benchmarking

  3. Assistant Professor

    09/2024 – 02/2025

    Universitat Politècnica de Catalunya

    Ran lab sessions and supervised undergraduate students through their digital systems design projects, mentoring them and giving technical guidance along the way.

    • Digital systems design
    • Lab sessions
    • Mentoring

  4. Developer Analyst

    09/2023 – 09/2024

    Deloitte

    Built and optimized solutions for major clients in the energy sector, with a focus on making bulk data operations faster and more efficient.

    • Energy sector
    • Bulk data operations
    • Performance optimization

  5. ML Research Engineer

    06/2022 – 08/2023

    FLAIR / UPC, EU Consortium

    Led research on Federated Learning and VEDLIoT integration in an EU-funded consortium: set up a 5G network, published open-source resources and optimized voice recognition algorithms across distributed federated deployments.

    • Federated Learning
    • 5G
    • VEDLIoT
    • Open source

Selected projects

EduCast: actual versus predicted enrolment heatmap
01 · 2026Featured

EduCast: Hierarchical Forecasting for University Enrollment

My Master's thesis on predicting course enrollment using student-level trajectory modelling. We compared LSTMs, GRUs, LightGBM, and Course2Vec embeddings on 505 students and 51 courses over 11 years, and explored hierarchical reconciliation to make forecasts consistent across different planning levels.

  • Hierarchical Forecasting
  • LSTM
  • LightGBM
  • Course2Vec
Read the case study
All projects

Stack

Tools I've worked with across research, production ML and full-stack engineering.

See them by area
PyTorchDeep Learning
TensorFlowDeep Learning
Scikit-learnML Library
NumPyScientific
PandasData Analysis
Hugging FaceNLP / LLMs
LangGraphAgent Framework
Google ADKAgent Dev Kit
n8nWorkflow Automation
Bedrock AgentsAWS AI Agents
PythonLanguage
CLanguage
SwiftLanguage
TypeScriptLanguage
LaTeXTypesetting
Google CloudGCP
AWSCloud
DockerContainers
GitVersion Control
LinuxOS
Next.jsWeb Framework
FastAPIAPI
PostgreSQLDatabase
RedisCache / DB
Neo4jGraph DB

Education

  1. MSc Machine Learning & Cybersecurity

    2023 – 2025

    Universitat Politècnica de Catalunya (UPC)MERIT Scholarship

  2. BSc Computer Engineering

    2018 – 2023

    EPSEM, Universitat Politècnica de CatalunyaEPSEM Scholarship

Let's connect.

Whether you'd like to collaborate, discuss machine learning, or simply exchange ideas, feel free to reach out.