Hi, I’m Ehsan
I’m a machine learning engineer and researcher with 5+ years in research and industry. I love how deep neural networks work inside, ever since a classmate presented GoogLeNet in undergrad and I was immediately captivated.
Since 2023, as an ML specialist at Pacefactory, I have built real-time perception for industrial cameras, from annotation (Grounding DINO, SAM) to model training and production deployment.
I did my MSc in Applied AI at the University of Ottawa (GPA 10/10, Vector Scholarship), researching generative models and adversarial attacks on face verification. That work became 3 peer-reviewed papers (IEEE, PMLR, MDPI) and 1 preprint.
Selected side projects
Masterface AttacksNine embeddings impersonate 47% of identities
LLM Face VerificationLoRA-tuned a 2B multimodal LLM to 98.7%; ResNet-50 still wins
ArcFace vs InfoNCEInfoNCE reaches 99.55% on LFW vs ArcFace's 99.72%
Trio DiffusionInfinite images, one patch at a time
Every car here is driven by a 16k-parameter diffusion policy running in your browser. You can drag any car or box around.
Email GitHub LinkedIn Scholar Medium Publications More projects