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

🐍 Contribution Snake

🚀 About Me

🧠 Research → production. I read the papers, then make the thing actually run.
🏭 Real systems: industrial visual inspection, real-time multi-stream video intelligence, recommendation, multimodal search.
⚡ The unglamorous parts: modular pipelines, VRAM-safe inference, hardware-aware deployment, honest evaluation metrics.
🎯 Now: publishing my portfolio here, one carefully documented repository at a time.

🧠 Neural Skill Atlas

_{How my technical mind is organised — 200 paths of real neuroanatomy, pulsing like a living brain.
Anatomy: James.mcd.nz · Hguiney, CC-BY-SA 3.0}

🛠️ Tech Stack

🧠 Deep Learning

👁️ Computer Vision

📊 Machine Learning & Data

🚀 Deployment & Practices

_{Modular pipelines · config-driven training · reproducible research · VRAM-safe inference · real evaluation metrics}

📌 Featured Work

📂 Open — 8 published projects, work in progress, and every public repo

🧭 code-context-engine ·

MCP context engine for AI coding agents — a local code graph answers "which files does this task need?". Recall 1.000 on repos it was never tuned on, 45% fewer context tokens in a real Claude Code session.
Rust MCP tree-sitter local-first

🔥 FireGuard — fire & smoke detection ·

Real-time fire and smoke detection for indoor CCTV with hazard-state alerting and instant email alarms — dataset audit → training → TensorRT export → live inference.
YOLO TensorRT OpenCV

🏴‍☠️ Concealed Face Detection ·

Tells a robber in a balaclava from a customer in a surgical mask, in real time — pose-based orientation, occlusion-hardened tracking, zero training.
InsightFace Pose Tracking

🪧 PelakX — license plate recognition ·

Open-source multilingual ALPR: vehicle + plate detection, Persian & Latin OCR, per-country plate grammar validation and temporal fusion. Iranian plates out of the box.
YOLO OCR Temporal fusion

⛑️ PPE Sentinel ·

PPE compliance monitoring for construction CCTV — person-grounded two-stage pipeline with ByteTrack, geometric head/torso association and temporal hysteresis.
YOLO ByteTrack TensorRT FP16

🧠 Visual Intelligence Engine ·

Multimodal search over an image archive — face recognition, species and food recognition, and open-vocabulary text search on one unified index.
CLIP InsightFace SQLite

🔬 PCB Component Classifier · Detection — YOLOv8

Explainable Random Forest over 33 hand-engineered CV features (93.3% holdout, every prediction traceable), plus a YOLOv8 baseline on 675k real annotations.
scikit-learn YOLOv8 OpenCV

📈 3D Point Cloud Curve Reconstruction

Benchmark of 5 peer-reviewed algorithms for denoising and curve reconstruction on industrial 3D-printing scans — 43 tests, CI green.
PCA Clustering Least-squares

🛠️ In preparation — VisionGuard, video surveillance platform, and more

VisionGuard — deep-learning AOI for surface defects and missing components on the production line. YOLO PyTorch
Intelligent Video Surveillance Platform — real-time AI over 32+ CCTV streams: face recognition, tracking, 14-class anomaly detection, fire/intrusion alerts. VideoMAE ArcFace
Shoplifting Behavior Detection — temporal video understanding with pose estimation and tracking. Video Transformers
Hybrid Freelancer Recommender — content + collaborative hybrid solving cold start, evaluated with Precision@K and NDCG. Autoencoder SVD
Medical Image Analysis — retinal disease classification and feature extraction. CNN Transfer learning

🗂️ Every public repository — auto-sorted by field, refreshed daily

🤖 AI & Computer Vision · 11 repos

code-context-engine · ⭐ 5 · _Rust
_{MCP context engine for AI coding agents}
PPE_detection-Sentinel · ⭐ 0 · _{Jupyter Notebook}
_{Real-time PPE compliance monitoring for construction CCTV — person-grounded two-stage pipeline (COCO person +…}
FireGuuard_Fire_detection_smoke · ⭐ 6 · _{Jupyter Notebook}
_{Real-time fire & smoke detection for indoor CCTV — YOLO26 detector, hazard-state alerting, and instant email…}
pelakx-license-plate-detection · ⭐ 4 · _{Jupyter Notebook}
_{Open-source multilingual ALPR / car license plate recognition}
thief-face-detection · ⭐ 5 · _{Jupyter Notebook}
_{Tells a robber in a balaclava from a customer in a surgical mask — real-time concealed-face detection for…}
melkai-realestate-platform · ⭐ 4 · _TypeScript
_{AI-powered Persian real estate platform — Next.js 16 demo (private preview)}
point-cloud-curve-reconstruction · ⭐ 5 · _{Jupyter Notebook}
_{Five peer-reviewed algorithms for 3D point cloud denoising and curve reconstruction, benchmarked on industrial…}
pcb-component-classifier · ⭐ 4 · _{Jupyter Notebook}
_{Explainable classification of electronic components from images - Random Forest over 33 hand-engineered CV…}
pcb-component-detection-yolov8 · ⭐ 5 · _{Jupyter Notebook}
_{YOLOv8 detection of electronic components on populated PCBs - 50 classes, 675k annotations, with dataset…}
visual-intelligence-engine · ⭐ 4 · _{Jupyter Notebook}
_{Multimodal search over an image archive - face recognition, species detection, food recognition and…}
ai-project-template · ⭐ 4 · _Python
_{Clean, reproducible project template for Computer Vision & Machine Learning - config-driven training, seeded…}

🌐 Web & Full-Stack · 3 repos

ParsaVictor.github.io · ⭐ 3 · _HTML
_{Mohammad Parsa Karkooti — AI Engineer portfolio (GitHub Pages mirror)}
portfolio · ⭐ 4 · _TypeScript
_{Motion-led personal portfolio — a WebGL particle instrument that morphs through the work as you scroll}
b2b-marketplace · ⭐ 5 · _TypeScript
_{B2B International Marketplace}

📊 GitHub Activity

_{Self-hosted and drawn from the GitHub API by my own Actions every day — no third-party stat cards.}

🧊 Isometric 3D Calendar

🌆 Skylines & City

_{Every tower is a day of my real contribution calendar. Rebuilt daily.}

_{My GithubCity by honzaap — every commit grows a building. Recorded weekly.}

🧱 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 extension — outputs a real .stl file:

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

🕒 Live Feed

⚡ What I shipped lately — recent activity and releases

Recent activity

⬆️ Pushed 1 commit to ParsaVictor/code-context-engine _{· 2026-10-01}
🔀 Opened a PR in ParsaVictor/code-context-engine _{· 2026-10-01}
🔀 Merged a PR in ParsaVictor/code-context-engine _{· 2026-10-01}
⬆️ Pushed 8 commits to ParsaVictor/code-context-engine _{· 2026-10-01}
⬆️ Pushed 1 commit to ParsaVictor/PPE_detection-Sentinel _{· 2026-09-27}

Latest releases

📦 code-context-engine v1.1.0 _{· 2026-10-01}
📦 code-context-engine v1.0.0 — v1.0.0 — the right files, measured on repositories it never saw _{· 2026-09-22}

📝 Say Hi & Guestbook

Passing through? Leave a note — I read every one, and it shows up on the wall above.

🤝 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