Skip to content
View mat21mf's full-sized avatar
5 followers · 9 following

Highlights

  • Pro

Block or report mat21mf

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. Markdown is supported. 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
mat21mf/README.md

Matias Rebolledo

Data and platform engineer based in Barcelona. I build the pipelines, infrastructure and validation tooling that scientists, statisticians and analysts rely on, with geospatial and scientific data as my specialty.

What I work on

  • Data platforms - batch pipelines and data lake architecture (Spark, Airflow, Trino, MinIO, Apache Iceberg), metadata and lineage (OpenMetadata), DuckDB for lightweight registries.
  • Geospatial and scientific data - GDAL, Rasterio, GeoPandas, xarray, PostGIS; GeoTIFF/COG, NetCDF, Parquet, STAC.
  • HPC and deployment - Slurm, Ansible, Docker/Singularity, GitLab CI/CD, Linux automation.
  • Validation services - FastAPI services and QC gates that stop bad data before it reaches downstream users.

Background

13+ years across quantitative analysis, official statistics and scientific computing. I started in statistics and econometrics, moved into geospatial data production and national-scale data lake engineering, and now work on climate model data pipelines (CMIP) at the Barcelona Supercomputing Center.

Currently exploring

  • Spatio-temporal inference on noisy data: GPS probe traces, map-matching, speed and closure estimation.
  • Decision rules under uncertainty: ranking competing actions or alerts by probability x consequence.
  • Combinatorial designs and their computational side (enumeration, search, GPU/HPC performance).
  • A typed systems language for production services alongside Python.

Selected work

  • probe-traffic-lab - noisy GPS probes to per-segment speeds and closures: HMM map-matching and estimation prototyped in Python, production core in a typed language, evaluated against known ground truth.
  • sandbox-rag-mcp - RAG scaffold for restricted environments: offline embedding, prebuilt LanceDB index, HF-free ONNX query path, and an MCP server that orchestrates a GPU workstation.
  • gdal-parquet - reproducible source build of GDAL 3.9 with GeoParquet, PROJ, GEOS and TileDB.
  • quiz_spaceag - Sentinel-2 NDVI time series for 33 avocado plots (2016-2020): shell ETL into per-plot daily grids, GeoPandas analysis and visualization.

Most of my professional work lives in institutional repositories.

Contact: LinkedIn | el4ur22lh@mozmail.com

Pinned Loading

  1. quiz_spaceag quiz_spaceag Public

    Quiz and solution regarding Sentinel Satellite Image Data

    Shell

  2. gdal-parquet gdal-parquet Public

    Shell

  3. probe-traffic-lab probe-traffic-lab Public

    Python