Skip to content
Perceptronix.
UK · Founded 2018 · Working globally

Production-grade AI for time-series forecasting and classification.

Perceptronix designs, builds, and operates machine-learning systems that make confident decisions in volatile, real-world environments — from macroeconomic forecasting with NIESR to production ensemble systems for capital markets and sports analytics.

5mo

earlier than the SPF
at detecting US inflation turning points

±0.2%

UK CPI inflation forecast accuracy across multi-month horizons

35+

peer-reviewed publications across AI, finance, and cognitive science

20+

years building neural systems for industry, academia, and capital markets

Illustrative

Inflation turning-point detection

MRN ensemble vs Survey of Professional Forecasters · USA CPI

MRN ensemble detects an inflation turning point five months earlier than the Survey of Professional Forecasters 6% 5% 4% 3% Jan Apr Jul Oct Jan Apr MRN signal SPF signal 5 months earlier
Realised CPI MRN ensemble forecast SPF benchmark
How we work

An end-to-end AI pipeline — not a notebook.

We replaced the legacy CRISP-DM lifecycle with a modern, production-first delivery model that mirrors how the best AI teams in industry actually ship.

01

Ingest

APIs · streams · DBs · FRED · Finage · OANDA · MetaTrader

02

Engineer

Cleaning · alignment · resampling · feature stores · lookback windows

03

Model

MRN ensembles · LightGBM · LLMs / RAG · agentic workflows

04

Deploy

GitHub Actions CI/CD · Docker · AWS EC2/EBS · Cloudflare

05

Monitor

Drift detection · walk-forward validation · stability dashboards

Decision impact: every step is wired end-to-end. We do not deliver a notebook — we deliver a continuously-running system you can trust on Monday morning.

Tooling

A modern stack — chosen for the job, not the hype.

Python PyTorch TensorFlow scikit-learn LightGBM XGBoost Pandas NumPy MariaDB PostgreSQL AWS EC2/EBS Cloudflare GitHub Actions Docker MLflow LangChain OpenAI Hugging Face
Let's build something measurable

Have a forecasting or classification problem that needs to work in production?

Tell us about it. First conversations are confidential, no-obligation, and usually end with a clear view of feasibility, data needs, and time-to-value.