AI research · systems engineering

Agentic intelligence,
built for the real world.

I’m Mahdi Mohseni, an AI research engineer and PhD researcher designing agentic, multi-agent, language, and multimodal systems that move from research questions to working products.

Explore my work Download CV
  • PhDMedical AI research
  • 4+ yearsIndustry experience
  • UKBased in the United Kingdom
Mahdi Mohseni near historic sailing ships

Currently exploring Autonomous systems for multimodal clinical data

Scroll to explore

01 · Profile

Research depth.
Product instinct.

My work sits where intelligent models meet dependable software systems.

I develop applied AI from first principles through deployment: shaping the problem, selecting and adapting models, designing tool and retrieval workflows, evaluating behaviour, and building the surrounding system that makes the result useful.

My current doctoral research at Ulster University focuses on agentic architectures for multimodal medical data integration and clinical decision support. Earlier industry work spans generative video, visual AI, and production computer-vision systems.

Download full CV Email me

02 · Core expertise

Systems that can reason,
act, and adapt.

I work across the model and systems layers, with a focus on AI that operates purposefully in real workflows.

01

Agentic & multi-agent systems

Architectures that coordinate specialised agents, tools, memory, and feedback for complex real-world tasks.

  • Tool use
  • Orchestration
  • Evaluation
  • Autonomy
02

LLMs, adaptation & RAG

Language systems grounded in the right context, adapted to the task, and assessed beyond a convincing demo.

  • Fine-tuning
  • PEFT / LoRA
  • Retrieval
  • RLHF
03

Multimodal & generative AI

Systems that connect language, images, video, and structured signals to solve applied research and product problems.

  • Vision
  • Diffusion
  • Video
  • Medical AI

03 · Selected work

Ideas made
operational.

Selected product and research work across intelligent documents, medical AI, video generation, and visual systems.

Flagship project · Active development

Owner · Founder · Creator · Developer

Turquoise
Jackal

A focused browser workspace for writing, compiling, reviewing, and sharing LaTeX projects—with intelligent, review-first assistance built into the editor.

I designed and developed the platform and continue to improve its capabilities, bringing the moving parts of serious academic writing into one clear workspace.

  • 01 Write and compile complete LaTeX projects
  • 02 Review precise lines and collaborate with clear roles
  • 03 Draft, revise, and repair LaTeX with Jackal AI
Visit Turquoise Jackal
A stylised preview of the Turquoise Jackal LaTeX workspace.
Research

Agentic AI for clinical data

Designing agentic architectures for multimodal medical data integration and clinical decision support as a PhD researcher at Ulster University.

  • Agents
  • Multimodal data
  • Federated learning
Industry

Generative AI for video

Led and contributed to research in video generation, image and video processing, and production-oriented deep-learning systems at Zebracat.

  • Video generation
  • Deep learning
  • Vision
Industry

Production visual AI

Developed systems for virtual staging, image enhancement, object removal, and automated design at Revivoto / AIhomeDesign.

  • Diffusion
  • Image editing
  • Optimisation

04 · Journey

Built through
research and practice.

A path through applied AI teams, independent engineering, and academic research.

Experience

  1. AI Research Engineer

    Zebracat · Berlin, remote

    Generative AI research for video, with work across deep learning, computer vision, and image and video processing.

  2. AI Research Engineer

    Revivoto / AIhomeDesign · Vancouver, remote

    Production visual AI for real-estate photography, including diffusion-based editing and inference optimisation.

  3. Machine Learning Engineer

    Independent · Isfahan

    Custom machine-learning systems and technical guidance for projects across different domains.

Education

  1. PhD, Medical Artificial Intelligence

    Ulster University · United Kingdom

    Agentic AI architectures for multimodal medical data integration and clinical decision support.

  2. MSc, Telecommunication Engineering

    Shahid Beheshti University · Iran

    Sensor fusion for 3D object detection in autonomous vehicles; 3.65 / 4.0 GPA.

  3. BSc, Electrical Engineering

    University of Isfahan · Iran

    Classical machine-learning and deep-learning approaches to handwriting recognition.

Working toolkit

Models & methods

LLMs · RAG · PEFT · LoRA · RLHF · diffusion · transformers · multimodal learning

Engineering

Python · PyTorch · Hugging Face · FastAPI · OpenCV · Docker · Linux · Git

Applied domains

Agentic systems · computer vision · generative video · medical AI · 3D vision

05 · Contact

Let’s build something
genuinely useful.

If you’re exploring an applied AI product, research collaboration, or an ambitious agentic system, tell me what you’re working on.

Emailmahdimohseni0333@gmail.com LinkedInConnect professionally GitHubExplore my code
Project enquiry

Open to thoughtful conversations

A little context helps — 20 characters minimum.

Your details are used only to reply to this enquiry.