Gustavo de Paula Avelar

Search and AI engineering, built on large-scale data systems.

I’m a senior software engineer at Envato in Melbourne, working on search. I came to it through data engineering and research: first distributed data platforms and clustering at scale on Spark and Hadoop, later a master’s on anomaly detection with ensembles. That’s still how I think about retrieval.

Work

Search engineering

I work on search infrastructure at Envato. Most recently I scaled our indexing system, part of a company-level goal to let other teams ship customer-facing features faster.

AI engineering

I build with LLMs, mostly agent-shaped tools. Loadstone is an engineering assistant I wrote in Rails on top of Claude: a registry of slash-command skills that pull context from Jira, GitHub, Slack and Confluence, plus background investigations that shell out to the Claude Code CLI so an agent can explore real repositories. Open source under Apache 2.0.

Publications

  1. 2021
    Characterizing and understanding ensemble-based anomaly detection

    G. de P. Avelar, G. O. Campos, W. Meira Jr.

    IX Symposium on Knowledge Discovery, Mining and Learning (KDMiLe 2021), pp. 153–160

    PDF DOI

  2. 2018
    Scalable and efficient data analytics and mining with Lemonade

    W. Santos, G. de P. Avelar, M. Horta Ribeiro, D. O. Guedes, W. Meira Jr.

    Proceedings of the VLDB Endowment, 11(12), pp. 2070–2073

    PDF DOI

  3. 2017
    Lemonade: a scalable and efficient Spark-based platform for data analytics

    W. dos Santos, L. F. M. Carvalho, G. de P. Avelar, Á. Silva Jr., L. M. Ponce, D. O. Guedes, W. Meira Jr.

    IEEE/ACM CCGrid 2017, pp. 745–748

    DOI

  4. 2017
    Comparação entre abordagens escaláveis para o processamento de conjuntos de dados textuais

    G. de P. Avelar, M. C. Naldi

    Revista de Informática Teórica e Aplicada, 24(1), pp. 121–149

    DOI

Background

My undergraduate work at UFV was on distributed data platforms and clustering at scale, using Spark and Hadoop. My MSc in Computer Science at UFMG was on anomaly detection with ensembles.

Alongside the MSc I worked on Lemonade, a Spark-based platform for data analytics built in the EuBra-BIGSEA collaboration. The papers above came out of all three.