ZEVIQ

AI agents for systematic trading · est. 2026

AI agents that research alpha.

Zeviq builds AI agents, and the platform they run on, to research, test, and run systematic strategies across futures, equities, options, and perpetuals.

held-out
Fig. 1 — Illustrative: synthetic data, not results.

01 — What we build

Agents. Platform. Strategies.

Research agents, the software platform under them, and the systematic strategies they produce, each informing the next.

  • AI Research Agents

    Agents that read research, form testable hypotheses, run experiments, and review the results. Every idea begins as a hypothesis, and most don't survive the process.

  • Alpha Research Platform

    Data pipelines, a grounded knowledge system, backtesting, and evaluation: the software our agents use to develop and validate alpha, built in-house.

  • Systematic Strategies

    Strategies that survive validation will run in deterministic code under hard risk limits, starting with equity-index futures.

02 — AI-native by design

AI researches. Code executes.

AI agents do the research, review, and monitoring. Trading decisions and risk limits will run in deterministic, tested code. Agents will be able to halt trading, never loosen a limit.

  • Agentic Research

    AI agents read our research library, turn ideas into testable hypotheses, and review code and results. Every idea stays a hypothesis until our own tests validate it.

  • Grounded Knowledge

    A retrieval system over curated research. Every answer cites its source, citations are checked by code, and when the evidence isn't there the system says so instead of guessing.

  • Evaluation-Driven

    Models and agents are measured before they're trusted: independent test sets, held-out validation, and latency and cost tracked for every operation.

03 — Markets

One platform, four markets.

  • Futures

    Equity indices first; metals, energy, rates

  • Equities

    Systematic and event-driven research

  • Options

    Volatility and hedging research

  • Perpetuals

    Crypto perpetual futures

04 — Philosophy

Built on evidence — not intuition, not narratives, not noise.

Markets do not reward conviction. They reward edge.

Every strategy we research starts as a hypothesis. We test against historical data and validate out-of-sample. Most ideas do not make it through.

Short-term performance is easy to manufacture. We focus on strategies that hold up across time and market regimes, and on agents whose work can be checked: every claim cited, every result reproducible.

05 — Careers

Work on hard problems.

Zeviq builds AI agents and research software for systematic trading. The work is research-driven and technically demanding. If that appeals to you, we want to hear from you.

See open roles →