Research timeline

2026
├─ Jan          Presentations continue, ~15 stakeholders to date.
2025
├─ Dec          Presentations continue.
├─ Late Nov     Demo 3, customer-centric on level crossings, presented to SNCF Réseau.
├─ Aug–Nov      Production platform built ("factory").
├─ Jun          Theoretical grounding via grant proposals.
└─ S1           Demo 2, adjusted to rail's specifics.
2024
├─ Nov–Dec      Demo 1, extra muros. Rail picked up.
└─ Q1–Q3        Prototyping: sources, analytics, substrate.

1. FY2025

The year started where 2024 left off. Early feedback from the rail sector on the November–December 2024 demo had started to come in.

1.1 2025-S1, demo adjusted to rail's specifics

The first three months of 2025 went into working through that feedback. We adjusted the analytical assumptions, how we presented results, and what the demo could and could not do. The result was an adapted version of the November–December 2024 demonstrator, tuned to the specifics of the rail sector. We took it through several rounds of presentations across the first half of the year, refining the rail-specific framing each time.

The pattern that emerged from those exchanges fed the grant proposals and applications written through the spring. Those documents, finalised around June 2025, first laid out in writing the theoretical basis of what we would ship in the autumn.

1.2 2025-Q4, customer-centric demo on one niche topic

The real platform came together between August and November. We refer to it internally as our "factory". The factory pulls in raw geospatial feeds and turns them into the per-asset and portfolio metrics that customers actually consume. The autumn release was designed on the back of the June 2025 theoretical grounding and the scope distilled from the S1 exchanges. From the broader candidate set, we picked one asset class to address end-to-end, the level crossings.

The first end-to-end production-grade run characterised the level crossings. We packaged it and presented it to SNCF Réseau in late November 2025. The model covers every active level crossing on the French national network across roughly 200 geospatial dimensions. The output reaches the user through two views. A portfolio view supports filtering across the whole network. A per-crossing view exposes origin–destination pair metrics.

Across 2025, the work was presented to roughly fifteen stakeholders across SNCF Réseau and other entities, through a mix of guided demos and self-served access, including rounds in December 2025 and January 2026.

1.3 External activity

Across the year, the team attended industry conferences and trade shows and ran prospection meetings and product demonstrations. The year-by-year list lives on the home page news section.

2. FY2024

The first full year of SAMRoute turned on choosing. We chose the data sources to build on. We chose the analytical approach to commit to. We chose the engineering substrate to lay down.

2.1 Prototyping and first demonstrator

The first three quarters of 2024 went into prototyping. The explicit goal was to produce a demo-able software that we could show to stakeholders of the infrastructure. Those choices converged in a first demonstrator shown in November–December 2024. It was designed extra muros of the sector. By definition, it ran disconnected from field feedback and field needs, and that gap was our first entry card, built to convey the analytical logic before pretending to production. We took it broadly to the public sector, road, and rail to gather feedback that would shape the next year's work. Rail picked up.

2.2 Platform foundation

The architectural blueprint that came out of this period (how raw data turns into results) lives on the home page. FY2024 settled the substrate underneath that chain. The substrate ingests data from OpenStreetMap, ESA Copernicus (GHSL and Worldcover), NASA GPW, IGN RGE Alti, Mapillary, ONISR, SIRENE, Eurostat, and GTFS. A prototype analytical pipeline sits on top.

3. R&D tax credit (CIR)

As a software editor, Oriskami SAS files its own annual CIR (Crédit d'Impôt Recherche) on the R&D it carries out on SAMRoute. The operations described in the year sections above document what those filings cover. The company also holds a CIR accreditation for consulting activities, first granted in 2022 and renewed in 2025 for the 2025–2027 cycle.

4. Ecosystem

Across 2024 and 2025 the work drew support from four organisations.

  • OVHCloud — Startup Program credits, February 2024 to February 2025.

  • CEREMA — multiple working meetings across 2024–2025. The public infrastructure-engineering reference in France whose materials we draw on.

  • SNCF Réseau — multi-touch engagement across 2025. Demos shown, level-crossing evaluation discussed, and the first production-grade run presented in late November 2025.

  • DGFiP — processes the company's annual CIR tax credit.