About me

Who am I?

I’m Mahdi Nasser: an engineer, AI builder and founder building practical AI systems for real business problems. I combine machine learning with an understanding of technical systems to build solutions that are meant to work in practice.
Through projects in document intelligence, LLM systems, computer vision, machine learning, signal processing and industrial automation, I have worked on turning complex data and manual processes into structured, usable solutions. I’m an electronics and automation engineer and am completing a Master in Cybernetics and Applied AI.
I started Nasser AI-Solutions to help businesses apply AI to real workflows, documents, data and systems. The goal is not impressive demos that never get used, but practical tools that save time, reduce manual work and provide a better basis for decisions.

How I work

My approach starts with the problem, not the technology. Before choosing a model, framework or method, I map the workflow, the data, the constraints and what the solution actually needs to improve.
Then I build AI systems that are testable, understandable and practical to use. I emphasize validation, structured output, error analysis and human oversight where needed, so AI doesn’t become a black box.

What I offer

I help businesses develop custom AI solutions that connect data, documents, processes and existing systems.
This can be automation of manual tasks, extracting information from PDFs and technical documents, AI tools on your own data, decision support, dashboards, computer vision or complete tools built around a specific workflow.
I combine AI models with the systems around them, with the goal of taking the solution beyond a prototype and towards a practical tool.

Values

Practical value
AI should solve an actual problem. I focus on solutions that save time, reduce manual work or make information easier to use.
Accuracy and validation
I emphasize testing, error analysis and quality assurance, so results can be assessed, improved and used with more confidence.
Security and control
Data, ownership and dependencies are considered early, so the solution is built with security and control as part of the design.
Custom over generic
The best AI solutions are built around the company’s actual workflow, data and needs, not around a one-size-fits-all solution that fits everyone poorly.

Technical expertise

  • Applied AI and machine learning:Machine learning, deep learning, NLP, LLMs, multimodal AI, model evaluation, error analysis and practical use of AI in concrete workflows.
  • Document intelligence and automation:Systems that extract, structure and quality-assure information from PDFs, documents, tables, images and technical diagrams.
  • Computer vision and multimodal AI:Image processing, OCR, visual patterns, object detection and analysis of data that combines text, images, video or speech.
  • From model to tool:Integration, data flow, interfaces and backend around the AI models, so they can be tested and used in real workflows.
  • Engineering and technical systems:Background in electronics, automation, embedded systems, signal processing, sensors, control engineering, PLC/VFD systems and motor control.

See how I build AI solutions in practice

See projectsContact me