Scientific Simulations

Scientific simulations in real notebooks.

Vicena gives the agent a persistent cloud JupyterLab for scientific Python. Ask for modeling, simulations, calculations, plots, fitting, or computational chemistry, and the agent builds the notebook with visible code and outputs. Selected advanced molecular modeling workflows are powered by Rowan through Vicena Compute.

JupyterLab

Python + SciPy

Rowan workflows

kinetics_model.ipynb
Vicena Scientific Simulations JupyterLab notebook interface

Agentic JupyterLab

Vicena creates notebooks, writes cells, runs them, reads output, fixes errors, and keeps the notebook visible so you can inspect every step.

Scientific Python stack

Use Python, NumPy, SciPy, pandas, matplotlib, RDKit, PySCF, ASE, Rowan SDK workflows, and installable packages inside the same persistent environment.

Rowan-powered modeling

Route selected advanced molecular modeling workflows through Vicena Compute while keeping inputs, workflow IDs, result tables, plots, and files in the notebook.

Persistent simulation state

Notebook files, generated figures, local datasets, Rowan workflow metadata, and installed packages stay with the Science Computer feature so later chats can continue from prior work.

What it is for

A simulation workspace for doing science, not just talking about it.

Use it for work that needs execution: notebooks, code, calculations, simulation, modeling, plots, and reproducible analysis. The agent can reason in chat, then move into Jupyter when the work needs code.

Model a system

Create a notebook for ODEs, reaction kinetics, transport, growth curves, population dynamics, or parameterized scientific models.

Run calculations

Do dimensional analysis, numerical integration, uncertainty propagation, fitting, optimization, statistics, and sanity checks with visible code.

Simulate and compare

Sweep parameters, generate plots, compare scenarios, inspect residuals, and iterate until the result matches the scientific question.

Model molecules

Submit bounded molecular modeling jobs, poll status, retrieve results, and turn Rowan outputs into plots, tables, and CSV files.

Molecular modeling

Rowan-powered chemistry workflows, inside Vicena notebooks.

Use Vicena Compute as the interface for advanced chemistry jobs: descriptors, pKa, conformers, tautomers, docking, molecular dynamics, orbitals, and related modeling workflows. The agent keeps the notebook visible, saves outputs, and respects computation-credit limits before submission.

Explore computational chemistry
Computed and reference ligand poses overlaid in a Vicena notebook

Gateway routed

Workflow IDs saved

Plots and CSV files

Notebook evidence

The notebook stays visible while the science runs.

Notebook cells, plots, figures, and generated artifacts stay in the foreground so the work can be inspected, extended, and challenged instead of disappearing into a chat summary.

CFD wake visualization

CFD wake visualization

Flow simulations and vortex/wake fields can be generated as notebook artifacts, then inspected with the model code beside them.

Graphene band structure

Graphene band structure

Interactive 3D electronic-structure plots can be generated from notebook code and inspected alongside the assumptions.

Docking pose ensemble

Docking pose ensemble

Rowan-powered molecular modeling outputs stay with the notebook as inspectable pose ensembles, figures, tables, and files.

Agent workflow

From question to executed notebook.

The important part is not that Vicena can write code. It can keep the code, the result, the error trace, the figure, and the scientific explanation in one reviewable notebook.

1

Plan

Define the scientific question, assumptions, inputs, and expected outputs.

2

Write

Create notebook cells with code, formulas, markdown notes, and plotting logic.

3

Run

Execute cells, inspect errors and outputs, then update the notebook in place.

4

Review

Summarize results, limitations, next experiments, and artifacts to preserve.

Example requests

Start with a scientific job, not a blank notebook.

Each request opens Vicena with a concrete notebook task. The agent should create the notebook, execute cells, and report what it learned from the outputs.