Six years ago, we were on the buy side. The tools available to us were either too slow, too generic, or too expensive to run at the latency modern markets require. So we built our own.
Nexara started in 2019 as an internal analytics tool at a mid-size systematic fund in Tallinn. We needed real-time signal generation, risk monitoring, and alternative data processing — and the existing vendors were charging enterprise prices for infrastructure that couldn't keep up with market microstructure changes.
By 2022, we'd rebuilt the core engine from scratch and started talking to other firms about the same problems. In 2023, we productized it. Now we deploy to banks, hedge funds, family offices, and asset managers who want institutional-grade AI without the institutional-grade implementation timeline.
We're based in Tallinn, Estonia. Remote-first, twelve people, entirely focused on one thing — building the best financial AI infrastructure available at any price point.
A signal that arrives 30 seconds late isn't a slow signal — it's a wrong signal. Everything we build is optimized for sub-200ms delivery because that's the only latency that matters in production.
Black boxes don't survive regulatory scrutiny and they don't earn analyst trust. Every model output comes with documented reasoning. If you can't explain a decision, you can't defend it.
We don't ask firms to replace their Bloomberg, their OMS, or their existing stack. We connect to it. The value is in the AI layer — not in forcing workflow disruption.
Twelve people, zero bloat. Every engineer at Nexara has shipped something that's in production. We stay small deliberately — it's how we stay fast.
Former portfolio manager at a systematic fund in London. 14 years in quant finance. Built the original Nexara risk engine in 2019 because nothing on the market worked the way the team needed it to.
ML engineer previously at Two Sigma and Citadel. Leads the signal generation and model infrastructure. 40+ published papers in applied ML for financial time series.
Ex-Goldman Sachs risk desk. Designed the real-time VaR engine and stress testing framework. Obsessed with tail risk and the scenarios models don't anticipate.
PhD applied mathematics. Leads alternative data research and the NLP pipeline for earnings intelligence. 6 years building sentiment models for fixed income and FX.
Platform infrastructure. Previously at Refinitiv building low-latency data distribution systems. Responsible for the 140ms average signal delivery that clients depend on.
Former buy-side technology consultant. Has onboarded 30+ institutional clients across Europe and North America. Speaks the language of both engineers and portfolio managers.
Small team. Hard problems. Real production scale.