Angel Robles
Portrait of Angel Robles

Angel L. Robles

Ecology · modeling software · AI-assisted scientific workflows

  • C++
  • R
  • SDM
  • Phylogenetics
  • HPC
  • Open source

I work at the intersection of ecology, evolution, and scientific computing. I develop machine-learning and mechanistic models for host–parasite interactions, climate-change biology, and species distributions.

My open-source work spans C++, R, Python, Bayesian inference, and HPC: from ecological niche and demographic SDMs to reproducible literature-mining and AI-assisted scientific workflows.

See a few selected projects below, the full project list, or my CV.

Selected projects

See all →

Modeling tools & methods

  • xsdm

    R

    An R toolkit for explainable species distribution modeling — fitting, evaluating, and interpreting SDMs in a reproducible workflow that plays well with modern tidyverse pipelines.

    Source Docs

  • xnicher

    R

    An R package for estimating ecological niche models with ellipsoidal environmental responses under an M hypothesis.

    Source

  • maxentcpp

    C++

    A header-light C++ implementation of MaxEnt for species distribution modeling. Designed to be fast on real ecological datasets and embeddable in larger pipelines.

    Source Docs

Data & web services

  • phyloseek

    Python

    A credible-set mammal phylogeny service — posterior-tree visualization and tree APIs for ecological and evolutionary analyses.

    Web service

  • genoseek

    Python

    A reproducible pipeline and queryable database mining the primary literature for species with population-genetic data, georeferenced samples, and reference genomes — for landscape-genetic and glacial-refugia studies. (Formerly genominer.)

    Source Web service

  • geoseek

    Python

    Frozen embeddings from an Earth-observation foundation model for ecology — landscape-similarity explorer and suitability demo.

    Source Demo

  • EcoSeek Bioclim

    Python

    ERA5-Land bioclimatic variables for ecological modeling, served through a reproducible web service.

    Source Web service

AI & scientific agents

  • WildSeek

    Python

    Autonomous wildlife disease surveillance and spillover forecasting — prospective host-susceptibility scoring from geographic, niche, and phylogenetic distances, literature-mined host–pathogen edges, and pre-registered predictions.

    Live monitor Source Dataset

  • EcoTagger

    Python

    Large-scale literature annotation for ecological interactions — tiered automated tagging with a human-validation benchmark (>100k verified records), a live interaction explorer, and a public triage/annotate API.

    Portal Extractor (HF Space)

  • Sciev

    Python

    System-One-style decision models on a frozen masked-diffusion backbone — typed questions answered with calibrated probabilities in a single forward pass, no text generation.

    Web service HF Space Source

  • EcoSeek

    Python

    A scientific agent environment for ecological and computational biology workflows. Domain-specialized LLMs, 30+ ecological tools, and a secure gateway for HPC clusters.

    Source About

  • EcoReasoner

    Python

    An experimental mixture-of-experts diffusion language model for scientific agents. EcoReasoner combines open, license-documented scientific corpora with tool-use traces to generate structured, repairable actions for ecological workflows such as biodiversity queries, climate-data retrieval, and species distribution modeling.

    Source Training dataset Beta model CoT traces

Get in touch

Reach me by email, or find me on GitHub and LinkedIn. I'm happy to talk about ecological modeling, scientific software, and reproducible HPC workflows.