Mohammad Parsa Karkooti ParsaVictor

_{640 dots in full colour
me → Python → OpenCV → neural net}

I design and ship end-to-end AI solutions — from problem definition and research to a working, optimized system running under real-world constraints.

مهندس هوش مصنوعی با تمرکز بر بینایی ماشین و یادگیری عمیق.
سیستم‌های هوشمند کاربردی را از تعریف مسئله تا استقرار نهایی طراحی و پیاده‌سازی می‌کنم.

⚡ Quick Navigation

🚀 About · 📈 Highlights · 🧠 Skill Atlas · 🛠️ Tech Stack · 🐍 Snake · 🧊 3D View · 🌆 Skylines · 📌 Featured Work · 🧭 Strengths · 📊 Stats · 🕒 Recent · 📦 Releases · 🎯 Pinned · 📝 Guestbook

🐍 Contribution Snake

_{Regenerated automatically twice a day by GitHub Actions.}

🚀 About Me

class ParsaKarkooti:
    role        = "AI Engineer"
    focus       = ["Computer Vision", "Deep Learning", "Video Intelligence"]
    domains     = ["Industrial AI", "Video Analytics", "Recommender Systems", "Multimodal AI"]
    education   = "B.Sc. Computer Engineering — Razi University, Kermanshah"
    approach    = "understand the problem → research the method → build it → make it fast"
    open_to     = ["AI Engineering roles", "Research collaboration", "Open source"]

🧠 I work at the intersection of research and production: reading the papers, then making the thing actually run.
🏭 Built real systems for industrial visual inspection, real-time multi-stream video intelligence, recommendation, and multimodal search.
⚡ Strong on the unglamorous parts — modular pipelines, VRAM-safe inference, hardware-aware deployment, real evaluation metrics.
🎯 Currently publishing my project portfolio here, one carefully documented repository at a time.

📈 Yearly Highlights

_{📊 Every year since I joined — straight from GitHub, refreshed daily}

_{Lifetime totals since 2024 · rebuilt daily at 03:00 UTC}

🔀 Branch Flow — last 7 days

gitGraph
   commit id: "quiet"
   branch thief-face-detection
   commit id: "09-04 · +3"
   commit id: "09-05 · +3"
   commit id: "09-06 · +3"
   commit id: "thief-face-detection · 8 commits this week"
   checkout main
   branch pelakx-license-plate-detection
   commit id: "09-04 · +2"
   commit id: "09-05 · +2"
   commit id: "09-06 · +2"
   commit id: "pelakx-license-plate-detection · 6 commits this week"
   checkout main
   commit id: "today"

🧠 Neural Skill Atlas

_{An anatomical map of how my technical mind is organised — 200 paths of real neuroanatomy, recoloured to the profile palette and pulsing like a living brain.}

🛠️ Tech Stack

Deep Learning

Computer Vision

Machine Learning & Data

Deployment & Practices

🧊 Isometric 3D Calendar

🌆 Skylines & Cities

_{Built from my real contribution calendar — every tower is a day, its height is that day's commit count. Rebuilt daily by Actions.}

Another angle — the same year as a solid block

Print it — the same graph as a real 3D-printable model, built with GitHub's official gh-skyline CLI extension:

gh extension install github/gh-skyline
gh skyline --user ParsaVictor --year 2026

Outputs a real .stl file — open it in any 3D viewer, or send it to a printer.

Drive it — the same graph as a drivable 3D city, courtesy of GithubCity by honzaap. Every commit grows a building.

Real screenshot of my GithubCity — buildings grown from actual contributions

📌 Featured Work

I'm publishing my portfolio here one repository at a time — each with architecture, reproducible setup, real metrics and a demo. Status badges below are honest and live.

PCB Component Classifier

Problem — inspection teams can't trust black-box AI on a production line.

Approach — Random Forest over 33 hand-engineered CV features; every prediction traceable to a readable decision path.

Result — 93.3% holdout accuracy, calibrated confidence lifts to 98.9% precision at 79% coverage.

scikit-learn XGBoost OpenCV

PCB Component Detection — YOLOv8

Problem — detecting 50 component classes on dense, populated PCBs from raw images alone.

Approach — YOLOv8 baseline on 675k real annotations, plus dataset tooling and an honest diagnosis of what limits the baseline.

Result — a documented reference point for what pure detection can and cannot do on PCBs.

YOLOv8 Ultralytics Roboflow

3D Point Cloud Curve Reconstruction

Problem — industrial 3D-printing scans are noisy; curve reconstruction must survive them.

Approach — benchmark of 5 peer-reviewed algorithms for denoising and feature-preserving reconstruction, with a full visual story.

Result — reproducible comparison with 43 tests, CI green, clear guidance on which method wins when.

PCA Clustering Least-squares

Concealed Face Detection

Problem — retail security needs to tell a masked robber from a customer wearing a surgical mask, in real time.

Approach — pre-trained models only, no custom training required, wired into a deployable pipeline.

Result — a practical, ethical-by-design concealed-face detector anyone can reproduce.

InsightFace OpenCV Real-time

VisionGuard

Industrial Visual Inspection ·

Deep-learning AOI system detecting surface defects and missing components on the production line.

YOLO PyTorch OpenCV

Intelligent Video Surveillance Platform

Modular real-time AI over 32+ CCTV streams — face recognition, tracking, anomaly detection (14 event classes), fire/smoke and intrusion alerts, per-camera web panel.

VideoMAE InsightFace/ArcFace SQLite

Shoplifting Behavior Detection

Temporal video understanding combining pose estimation and object tracking to flag suspicious in-store behavior.

Pose estimation Tracking Video Transformers

Hybrid Freelancer Recommender

Content + collaborative hybrid solving the cold-start problem, evaluated with Precision@K, Hit Rate and NDCG.

Deep Autoencoder TF-IDF Truncated SVD

Medical Image Analysis

Retinal disease detection: classification and feature extraction.

CNN Transfer learning

Open-Vocabulary Media Search

(built, public repo pending)

Natural-language search across a media library with face/species/food recognition and incremental embeddings, on a VRAM-safe pipeline.

CLIP ViT-L/14 (FP16) InsightFace SQLite

🧰 Open source

ai-project-template — a clean, reproducible starting point for CV/ML projects. Config-driven training, seeded runs, framework-free metrics, pytest suite and CI out of the box.

🗂️ Repository Index

_{🔄 Every public repo I own, sorted by field — this index grows automatically whenever I publish something new.}

🤖 AI & Computer Vision · 8 repos

Repo	What it is	Stars
`pelakx-license-plate-detection`	🚗 PelakX — Multi-country License Plate Intelligence (ALPR): real-time vehicle detection, tracking, script-aware OCR, plate-category & region recognition for any country's grammar — deepest support for Iran 🇮🇷 today (taxi/government/police/diplomatic/free-zone plates, province lookup), extensible to yours. YOLO + OCR + country grammar.	⭐ 1
`fireguard`	🔥 Real-time fire & smoke detection with hazard-state intelligence — 3 generations of YOLO (v8/11/26), broadcast-grade overlays, zero training. One-click Colab demo. [Demo release]	⭐ 1
`thief-face-detection`	Tells a robber in a balaclava from a customer in a surgical mask — real-time concealed-face detection for retail CCTV. Pose-based orientation, occlusion-hardened tracking, zero training.	⭐ 2
`melkai-realestate-platform`	AI-powered Persian real estate platform — Next.js 16 demo (private preview)	⭐ 1
`point-cloud-curve-reconstruction`	Five peer-reviewed algorithms for 3D point cloud denoising and curve reconstruction, benchmarked on industrial 3D-printing scan data	⭐ 2
`pcb-component-classifier`	Explainable classification of electronic components from images - Random Forest over 33 hand-engineered CV features, 93.3% holdout accuracy, every prediction traceable to a readable decision path	⭐ 1
`pcb-component-detection-yolov8`	YOLOv8 detection of electronic components on populated PCBs - 50 classes, 675k annotations, with dataset tooling and a full diagnosis of what limits the baseline	⭐ 1
`ai-project-template`	Clean, reproducible project template for Computer Vision & Machine Learning - config-driven training, seeded runs, tests and CI out of the box.	⭐ 1

🌐 Web & Full-Stack · 1 repo

Repo	What it is	Language	Stars
`b2b-marketplace`	B2B International Marketplace		⭐ 2

🧭 Strengths

Computer Vision

Object detection & tracking, face recognition, pose estimation, image retrieval, video understanding, open-vocabulary / multimodal models.

AI Systems Engineering

Modular inference pipelines, real-time multi-stream processing, DB-integrated AI apps, model optimization, hardware-aware deployment.

ML & Data Science

Feature engineering, EDA, classical ML (XGBoost/LightGBM/SVM), recommender systems, rigorous offline evaluation.

📊 GitHub Activity

🕒 Recent Activity

⬆️ Pushed 1 commit to ParsaVictor/pelakx-license-plate-detection
⬆️ Pushed 1 commit to ParsaVictor/thief-face-detection
🔹 Delete on ParsaVictor/pelakx-license-plate-detection
🔀 merged PR in ParsaVictor/pelakx-license-plate-detection
🔀 opened PR in ParsaVictor/pelakx-license-plate-detection

📦 Latest Releases

No releases published yet — this section will fill in automatically the day I ship one.

🎯 Pinned Repositories

📌 Whatever's actually pinned on the profile

_{🔄 Auto-refreshed daily · 2 pinned repos · language colours match GitHub's own}

Repo	Description	Language
`ai-project-template`	Clean, reproducible project template for Computer Vision & Machine Learning - config-driven training, seeded runs, tests and CI out of the box.		⭐ 1 · 🍴 0
`ParsaVictor`	My GitHub profile README - AI Engineer \| Computer Vision & Deep Learning		⭐ 3 · 🍴 0

[![Explore all repositories](https://img.shields.io/badge/%F0%9F%94%AD_Explore_all_repositories-github.com%2FParsaVictor%3Ftab%3Drepositories-F90001?style=for-the-badge&logo=github&logoColor=white)](https://github.com/ParsaVictor?tab=repositories)

📝 Say Hi & Guestbook

Passing through? Leave a note — I read every one. The guestbook lives in issues; it doubles as my feedback wall.

🤝 Let's build something

I'm open to AI Engineering roles, research collaboration, and interesting computer-vision problems.

"Turning technical knowledge into results that work in the real world."

Navigation Menu