◎
Multimodal AI
Combining visual and textual information with vision-language models, embeddings, and semantic understanding.
VLMs · Embeddings · Retrieval
AI ENGINEER / MILAN, ITALY
Building AI systems that turn complexity into clarity.
I'm Emad Bahreinipour — a computer scientist and AI engineer working across computer vision, multimodal models, and intelligent software systems.
From research questions to practical products.
FIELD NOTES / 001INTELLIGENT SYSTEMS ✳
INPUT → INSIGHTVisual reasoningsignals / semantics / context
01 PERCEPTION 02 REASONING 03 DECISION COMPUTER VISION↗MULTIMODAL AI
01 / WHAT I DO
Between research
and reality.
I enjoy the difficult part of AI engineering: connecting promising methods to systems that are useful, measurable, and maintainable.
◈
Computer vision
Analyzing images through learned representations, visual anomaly detection, and digital forensic techniques.
Visual analysis · Forensics
⌘
Applied AI systems
Building practical pipelines that connect models, services, data stores, and interpretable outputs.
Python · APIs · Vector DBs
02 / SELECTED WORK
Work worth exploring.
CASE STUDY 001RESEARCH / ENGINEERING
IMAGE INPUT
Visual evidence
SEMANTIC SIGNAL
FORENSIC SIGNAL
FLAGSHIP PROJECT2026
Making visual evidence understandable.
My master's thesis explored how vision-language models and digital forensic techniques can work together to examine insurance claim images and surface meaningful inconsistencies.
Explore case study
ENGINEERING LABIN DEVELOPMENT
Small experiments.
Big questions.
An evolving space for independent experiments in retrieval systems, applied AI, and building from first principles. Future work will be published with code and implementation notes.
05 / GET IN TOUCH
Let's make something
useful together.
Open to conversations about AI engineering, thoughtful projects, and interesting ideas.