Vamimi · CEO
Ahmed Hamimi
CEO, VamimiAI engineer and data scientist
Ahmed Hamimi is the CEO of Vamimi and leads its product and engineering. An AI engineer and data scientist, he designs the agentic system at the centre of VamimiAI: an assistant that understands a request, plans the work with tools and checks its own output before anything reaches the screen.
- Position
- CEO, Vamimi
- Discipline
- AI engineer and data scientist
- Leads
- Product and engineering
- Contact
- a.mhamimi@outlook.com
The company
About VamimiAI
VamimiAI is an AI assistant for visual thinking. Ask a question and it answers in words; ask for a diagram, a chart, a table or a timeline and it builds one, complete, laid out and ready to share. Each visual can then be refined by conversation or edited by hand.
It works from whatever the person brings: a sentence, notes from a meeting, a PDF, Word, Excel or CSV file, or an exported n8n, Make or Zapier workflow. Every result downloads as SVG, PNG, JPEG or PDF, and flow diagrams also export to draw.io and Mermaid.
Capabilities
What VamimiAI makes
Flow and process diagrams
Flowcharts, swimlanes and architecture diagrams, with every decision path drawn.
Sequence diagrams
Who sends what to whom, in order, for systems and conversations.
Charts, tables and KPI tiles
Numbers turned into a chart, a comparison table or a row of key figures.
Timelines
Plans, roadmaps and histories, laid out in order.
Mind maps and org charts
Ideas around a central topic and teams of any size, kept readable.
Documents and workflows
Attached files read as context, and automation exports imported exactly as they are.
The technology
How VamimiAI uses agentic AI
VamimiAI is built as an agent, not a one-shot generator. The model decides, plans and acts through tools, and code checks every step it takes.
Decides before it acts
For every message, VamimiAI first decides whether to answer in words, ask a question back or build something. It never draws what was not asked for.
Plans the work with tools
To build, the model writes a plan as a sequence of tool calls: add a step, connect two steps, group them, make a chart or a table. It never places anything by coordinates.
Checks every call
Each call is validated on its own. A slip with one obvious meaning, such as a mistyped reference, is corrected in code; a call that fails is reported without stopping the rest of the work.
Reviews its own result
Logic checks look for a decision with one exit, a step that leads nowhere and a step nothing reaches. Only the broken part goes back to the model for repair, and anything still open is shown as a note, never hidden.
Draws with a measured engine
A deterministic layout engine turns the plan into the picture and verifies the geometry: overlaps, text that does not fit, labels on top of each other. When something would be hidden, it lays the visual out again with more room.
Continues from the person's version
The assistant sees the current visual and changes only what is asked. Edits made by hand are described back to it in words, so the next turn carries on from the person's own version.
Expertise
Areas of focus
- Agentic AI systems
- Applied machine learning
- Data science and analysis
- Language model tooling
- Layout and visualisation engines
- Product engineering
Connect
Get in touch
- LinkedInProfessional profile and career history
- GitHubCode and engineering projects
- Emaila.mhamimi@outlook.com
For help with the app, billing or your data, see the contact page.
See the work for yourself.
Describe what you want to see, attach a document, or import a workflow, and refine the result by asking.
Open VamimiAI