English

Numerical simulation & applied AI

OptiIA Consulting helps businesses, laboratories and engineering teams connect physical models, experimental measurements and data. Our focus: optics, photonics, scientific signal analysis and Python-based tools.

Start with a focused study

Simulation study — Explore an optical or physical model, compare configurations and assess parameter sensitivity. Inputs: system description, parameters and available measurements. Deliverables may include an agreed model, scripts, figures and a technical report.

Scientific data & Python automation — Turn experimental files into a repeatable analysis workflow. Inputs: sample files, processing steps and expected outputs. Deliverables may include documented scripts, visualisations and operating instructions.

AI feasibility & prototype — Assess whether your signals or images can support a prediction, classification or detection task. Inputs: representative data, available labels and evaluation criteria. Deliverables may include a baseline method, a prototype evaluated on separate data and a report on performance and limitations.

Each engagement starts with an agreed scope, deliverables and validation criteria. Pricing and timelines are confirmed in a tailored proposal.

How we work

1. Scope the question, data and constraints.

2. Build a model or prototype and review intermediate results.

3. Validate against measurements or separate test cases and document assumptions, limitations and next steps.

Two reproducible examples

Gaussian beam propagation: an analytical free-space model at 1064 nm compares 100 and 200 µm beam waists. At 500 mm from the waist, beam radii are 1.696 and 0.870 mm. Assumptions: ideal M² = 1 beam and paraxial propagation; no apertures or aberrations.

Signal amplitude estimation: a Ridge regression trained on 500 synthetic noisy sinusoids is evaluated on 200 independent synthetic test signals. Mean absolute error: 0.0145 arbitrary units for Ridge versus 0.0130 for a conventional FFT baseline. The simpler method is slightly better in this example. Frequency is known and fixed; these results do not establish performance on industrial data.

These are demonstrations, not client projects. Figures and detailed assumptions are available on the Démonstrations page.

Your direct contact

Ali Yaghi, founder of OptiIA Consulting, is your contact throughout the study. His academic background includes the OPTIQ master’s programme at Université Côte d’Azur, with projects in optical simulation, laser signal analysis and applied machine learning.

Discuss your project — 20 minutes

Email contact@optiia-consulting.fr with your objective, available data and target timeline. A short summary is enough for an initial exploratory conversation; confidential files are not required at this stage.