Skip to content
AsteriaX Labs

About

A small studio with a wide field of view.

An independent collaboration connecting machine learning, cybersecurity, and infrastructure. Different disciplines. A shared commitment to understanding systems by building and studying them.

Different perspectives. A shared engineering instinct.

Federated intrusion detection, malware analysis, and renewable power for telecom sites look unrelated until you look at their common challenges: uneven data, uncertain operating conditions, and systems whose reliability depends on the decisions made beneath the surface.

Alongside the research, the studio builds software it actually runs, including an AI gateway and a self-hosted media platform. These projects create space to explore integration, infrastructure, and the practical decisions that connect an idea to a working system.

AsteriaX Labs is an independent, unincorporated research and engineering collaboration. It is not a registered corporation.

Principles we hold ourselves to.

  1. 01

    Baseline before breakthrough

    Every claim of improvement needs something honest to be measured against. We reproduce first, then change one thing at a time.

  2. 02

    Say where the work stands

    In progress, replicating, or still scoping: each direction carries its real stage, and nothing is promoted ahead of its evidence.

  3. 03

    Credit what we build on

    Published methods are named as external foundations. The contribution is the change we make, never the framework we start from.

  4. 04

    Build things that get used

    Engineering projects are operated, not demonstrated. They are kept distinct from research and described without inflation.

  5. 05

    Results when they exist

    No figures, benchmarks, or outcomes are published before the experiments behind them are complete and checked.

Collaborators.

  • Will

    Federated Learning & Cybersecurity

    Will's thesis research examines federated intrusion detection under non-IID data, working within PROTEAN, an existing prototype-based framework.

    The current investigation focuses on how the server weights client prototypes during aggregation, and what that choice means for rare attack classes.

    Leads

    Federated Learning & Cybersecurity

    Ongoing Research
    View direction
  • Yoga

    Malware Detection & Deep Learning

    Yoga's research concerns malware detection with deep representation learning and heuristic search for feature selection.

    Work is currently at the baseline stage: studying and replicating an external published approach before defining any thesis contribution.

    Leads

    Malware Detection & Deep Learning

    Baseline Investigation
    View direction
  • Faisal

    Renewable Energy & Telecommunications

    Faisal's research explores renewable power for telecommunications infrastructure, with particular interest in solar-assisted power systems for base transceiver stations.

    The direction is exploratory. The final research question and methodology have not yet been confirmed.

    Leads

    Renewable Energy & Telecommunications

    Exploratory Research
    View direction

Reach the people doing the work.