About us
We turn messy tick archives into reliable Forex datasets and deterministic M1–D1 OHLCV bars with audit-grade validation — ready for backtesting and AI/SMC workflows.
Who we are & what we do
cortex spectra blends data engineering, AI/SMC, and application development (desktop & web) to give you reliable data and measurable models. Our solutions are designed to handle real-world pressure: large tick archives, iterative research cycles, and strict validation conditions. We do not ship the “fastest possible” — we ship the most reliable that you can reproduce.
- Tick Data QC — detects duplicates, out-of-order timestamps, weekend ticks, unrealistic spreads, and gaps.
- Bar Generator — deterministic M1 bars and precise H1/H4/D1 aggregation that preserves OHLC extremes.
- SMC & AI — advanced features (FVG, OB, CHoCH, liquidity…), Transformer/LSTM models, walk-forward evaluation.
- Tooling — PySide6 desktop tools, ASP.NET Core sites, automations, validation reports, and CI/CD.
Quality & reproducibility
Same inputs → same outputs. Stable schema, clear filter rules, and step logs.
Performance without noise
Speed matters, but never at the expense of accuracy. We optimize where it counts.
Security & NDA
Isolated environment; we sign an NDA and share results only.
Transparency
Clear inputs/outputs, logs, and reports — so you can verify every decision.
How we work
- Scope — short exchange on symbols, periods, and formats.
- Plan — we propose a path (QC → bars → validation) and timelines.
- Build — run QC/bar generation, record steps and metrics.
- Validate — check HL-ratio, minute alignment, outliers, and schema stability.
- Deliver — Parquet and/or CSV.GZ + optional
*_stats.csvsummary.
Why us
- Engineering discipline — no QC shortcuts; no “magic” heuristics.
- Real-world focus — tools and datasets you can plug into work immediately.
- Iterative development — fast hypothesis testing without compromising data integrity.
Technologies & tools
| Stack | Description |
|---|---|
Python (pandas, pyarrow, TensorFlow) | Tick processing, bar generation, feature engineering, and AI models. |
ASP.NET Core + Bootstrap | Fast, optimized sites and panels to review results. |
PySide6 desktop | Practical GUI tools for QC and pipeline runs on client side. |
Git / CI + Colab | Reproducibility, experiment versioning, and scalable training. |
Parquet / CSV.GZ | Stable formats for long-term storage and fast analytics. |
We pick tools that maximize reliability per unit of time.
Related services
Explore our specialized services — from forensic QC to deterministic bar generation.
FAQ
Do you sign an NDA?
Yes. We sign an NDA before any data transfer and work in an isolated environment; we share results only.
Can I try with a small sample first?
Absolutely — send 1–2 files; we’ll provide a QC report and a small bar set to review.
Do you build custom web/desktop tools?
Yes — we build PySide6 desktop and ASP.NET Core web tools that match your process and validations.
Let’s make your data and models production-grade
Send symbols/months and target formats — we’ll propose a plan and deliverables.
Contact See services