Skip to content
View hammadhaideer's full-sized avatar
1 follower · 0 following

Block or report hammadhaideer

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
hammadhaideer/README.md

Hey, I'm Hammad 👋

M.S. Computer Science | Xinjiang University, China  |  Visual Anomaly Detection | PEFT | Foundation Models

I'm a Pakistani CS researcher based in Ürümqi, China, doing my master's at Xinjiang University. My research is in visual anomaly detection, teaching models to spot defects and anomalies in images across industrial, logical, and medical domains.

Right now I'm working on combining continual learning and parameter-efficient fine-tuning (LoRA, adapters, visual prompts) on vision foundation models like DINOv2 and CLIP. The goal is to make models that can keep adapting to new domains without forgetting what they already know.

Before the master's I fine-tuned BERT for production NLP at Bytewise and built an end-to-end AI platform as my final year project which got me a 4.0/4.0 for it, which I'm still pretty happy about.

Stack

Engineering: Python · PyTorch · TensorFlow · Hugging Face · NumPy · OpenCV · FAISS · Docker · FastAPI · GitHub Actions · Linux · Git
Research: DINOv2 · CLIP · SAM · LoRA · PEFT · PatchCore · EfficientAD · WinCLIP · AnomalyCLIP · UniVAD · MVTec-AD · MVTec-LOCO · VisA
Data: SQL · MongoDB · AWS S3

Research: DINOv2 | CLIP | SAM | LoRA | PEFT | PatchCore | EfficientAD | WinCLIP | MVTec-AD | VisA

Open to

  • Research internship Summer 2026
  • Paper co-authorship in VAD, continual learning, or PEFT
  • Just a good conversation about foundation models

Pinned Loading

  1. patchcore-reproduced patchcore-reproduced Public

    Clean reproduction of PatchCore (Roth et al., CVPR 2022) on MVTec-AD with image-level AUROC, pixel-level AUROC, and pixel-level AUPRO. First in a series of visual anomaly detection reproductions.

    Python

  2. winclip-reproduced winclip-reproduced Public

    Clean reproduction of WinCLIP (Jeong et al., CVPR 2023) on MVTec-AD and VisA — zero-shot and few-normal-shot anomaly classification and segmentation with CLIP.

    Python