An autonomous agent making decisions in a live market
Delivered by members of Exana's founding team in previous roles — anonymized to protect client confidentiality.
- AI & Automation
- Cloud Architecture
- System Integrations
Project: an Exana R&D project (autonomous trading agent)
Service demonstrated: AI & Automation · Cloud Architecture · System Integrations
The problem
Most AI demos run where mistakes are cheap. We wanted the opposite: an environment where a decision is acted on immediately, the data never stops moving, and an unnoticed failure costs something real. A live market is that environment — which makes it a hard test of whether an agent can be trusted to run unattended.
What was built
An autonomous agent in Python, connected to exchange APIs through CCXT, with strategy pipelines defined in configuration rather than hard-coded — so a rule can be changed without touching the engine. A FastAPI dashboard shows what the agent is doing in real time, and a Telegram bot pushes alerts when something needs a human. The whole thing is containerised with Docker so it runs the same way anywhere.
The outcome
A working demonstration of applied AI in a decision-critical, always-on setting: an agent that acts, a dashboard that shows why, and alerting that puts a person back in the loop when it matters. Those three parts are what we build into any autonomous system a client depends on.
Research and development only. This is not a financial product, and we make no claims about trading performance.
Still deciding
Start the brief anyway. It takes about two minutes, every question has a "not sure" answer, and you will get a written reply within one business day telling you which of these fits — including the answer that none of them does and here is what we would do instead.