Terminal • Projects • Architecture • Tech Stack • 30DaysOfAI • Milestones • Turkce Ozet

---

Professional Network & Profiles

---

Terminal Session

suleyman@devsecops-lab:~$ whoami
+-----------------------------------------------------------------------------+
|  Name: Suleyman Toklu                                                       |
|  Role: Security-Minded AI Engineer                                          |
|  Base: Isparta University of Applied Sciences (ISUBU), Computer Engineering |
|  Focus: External Recon Automation, Explainable Pentest AI, Quantum ML       |
+-----------------------------------------------------------------------------+

suleyman@devsecops-lab:~$ cat core_philosophy.json
{
  "mission": "Automate repetitive recon tasks, maintain human control over critical decisions, report with clarity.",
  "active_initiatives": [
    "pentest-cli -> Human-in-the-loop external penetration testing orchestrator",
    "smart-caliper -> Sub-pixel computer vision metrology and defect inspection",
    "qubit-architect-v2.0 -> Benchmark platform for Quantum ML vs Classical ML",
    "LLM-Comparison-Benchmarking -> Controlled evaluation of code generation models"
  ],
  "milestone": "30DaysOfAI -> 30 production AI applications shipped across 30 consecutive days"
}

---

Featured Projects

Status	Project	Focus & Technical Innovation	Stack	Repository
	pentest-cli	Human-in-the-loop web penetration testing orchestrator. Autonomous LLM justification before scans with strict approval gates.	Python Security LLM	Repository
	smart-caliper	Sub-pixel computer vision metrology and automated manufacturing defect inspection engine.	Python OpenCV Metrology	Repository
	QML	Comparative performance evaluation of quantum machine learning algorithms using Qiskit (BSc Thesis).	Qiskit Python QML	Repository
	qubit-architect-v2.0	Interactive web platform for benchmarking hybrid classical and quantum variational machine learning circuits.	TypeScript React Next.js	Repository
	30DaysOfAI	Consolidated monorepo of the 30-day AI marathon. 30 distinct ML, CV, NLP and Audio models deployed live to Streamlit and Hugging Face.	Python Streamlit Gradio	Repository
	LLM-Comparison-Benchmarking	Controlled benchmark comparing code-generation models on architectural constraints, contracts, and adversarial tests.	Next.js TypeScript Python	Repository
	image-processing	End-to-end 10-module Computer Vision curriculum from pixel math to Canny, Hough, and Haar cascades.	Python OpenCV NumPy	Repository
	deep-learning-foundations	Comprehensive Deep Learning curriculum: Multi-class Softmax, CNNs, MobileNetV2 Transfer Learning, and LSTMs.	TensorFlow Keras Python	Repository
	ml-foundations-journey	8-chapter Machine Learning engineering roadmap: EDA, XGBoost, Optuna tuning, PCA, and SMOTE for imbalanced data.	Scikit-Learn Optuna Python	Repository

---

Live CLI Execution Pattern — pentest-cli

$ pentest-cli --target target.example.com --scope authorized_domains.txt --mode assisted

[*] [PHASE 1: PASSIVE RECON]
    Running subfinder and asset discovery...
    [OK] 23 live subdomains discovered across target scope.

[*] [PHASE 2: ACTIVE PROBING]
    Running httpx and selective port telemetry...
    [OK] Active web services identified on ports 80, 443, 8080, 8443.

[*] [LLM AGENT RATIONALE]
    Identified legacy admin endpoint on staging.target.example.com:8443 with TLSv1.0.
    Recommended action: Execute selective configuration check template.
    Strict constraint: Zero aggressive fuzzing or denial-of-service payloads permitted.

[?] ACTION REQUIRED: Approve scan on target port 8443? [y/N]: y
    [>] Executing verified template bundle...
    [OK] Scan finalized. Zero unauthorized exploits executed.

[*] [EXPORT] Executive summary written to ./reports/2026-09-pentest-summary.md

---

Architecture — pentest-cli

%%{init: {'theme': 'neutral', 'themeVariables': {'fontFamily': 'ui-monospace, SFMono-Regular, Menlo, Monaco, Consolas, monospace'}}}%%
flowchart LR
    subgraph S1 [Phase 1: Authorization]
      A[Scope Target & Written Authorization]
    end

    subgraph S2 [Phase 2: Discovery Engine]
      B[Passive Reconnaissance<br/>subfinder & assetfinder]
      C[Active Probing<br/>httpx & nmap telemetry]
    end

    subgraph S3 [Phase 3: Vulnerability Intel]
      D[Vulnerability Assessment<br/>nuclei & configuration checks]
      P[LLM Agent Reasoning<br/>Tool Selection Rationale]
    end

    subgraph S4 [Phase 4: Human Gate & Report]
      H[Human-in-the-Loop Verification<br/>Operator Review & Approval]
      R[Executive Report Generation<br/>Findings & Remediation]
    end

    A --> B
    B --> C
    C --> D
    D --> P
    P --> H
    H -->|Approved| R
    H -->|Revise / Refine| P

Loading

---

The 30 Days of AI Sprint Spotlight

Between August and September 2025, I completed the 30 Days of AI Challenge: 30 distinct AI/ML applications built, trained, and deployed in 30 consecutive days.

Day	Project Name	Focus & Domain	Link
Day 30	AI Swiss Army Knife	Multi-tool AI suite integrating text, image, and tabular intelligence	Inspect Day 30
Day 29	AI-Supported Sniper Lab	Computer vision targeting analysis and trajectory prediction model	Inspect Day 29
Day 28	Transformer Explorer	Interactive attention map visualizer and transformer decoder explorer	Inspect Day 28
Day 26	AI Music Studio	Audio synthesis and neural music genre composition engine	Inspect Day 26
Day 23	AI Image Generator	Generative latent diffusion pipeline with interactive prompt tuning	Inspect Day 23
Day 14	YOLO Object Detection	High-throughput real-time object detection and spatial classification	Inspect Day 14

Explore all 30 production applications on the central monorepo: 30DaysOfAI

---

Tech Stack & Tooling

Languages & Core Scripting

AI / ML & Computational Science

Web & Applications

Security, Recon & Infrastructure

---

GitHub Metrics & Activity

---

Engineering Milestones

%%{init: {'theme': 'neutral', 'themeVariables': {'fontFamily': 'ui-monospace, SFMono-Regular, Menlo, Monaco, Consolas, monospace'}}}%%
timeline
    title Engineering and Research Timeline
    2022 : Foundational algorithms and systems programming in C#, C++, and Java
    2024 : Classical Machine Learning pipelines and data engineering
    2025 H1 : Applied Cryptography toolkits (AES-SHA256, ECC-512) and time series forecasting
    2025 Aug-Sep : 30DaysOfAI Sprint with 30 production AI applications shipped
    2025 Q4 : Quantum Machine Learning thesis research (QML) and combinatorial optimization
    2026 : qubit-architect-v2.0, smart-caliper, and LLM code benchmark evaluations
    Current Focus : pentest-cli and autonomous reconnaissance orchestrators

Loading

---

Turkce Profil Ozeti ve Hakkimda (Genisletmek icin tiklayin)

Hakkimda

Isparta Uygulamali Bilimler Universitesi Bilgisayar Muhendisligi ogrencisiyim. Yapay Zeka (AI/ML) ve Siber Guvenlik (DevSecOps & Pentest) alanlarinin kesisiminde calisiyorum.

Muhendislik yaklasimim: Tekrarlayan ve zaman alan surecleri otomatize etmek, karar aninda insan denetimini (Human-in-the-Loop) merkezde tutmak ve sonuclari acik ve net bir teknik dille sunmaktir.

One Cikan Calismalar:

1. pentest-cli: Harici web sizma testlerini asamali olarak yoneten, her tarama adiminda LLM destegiyle gerekce sunan ve operator onayi olmadan hicbir saldiri komutu calistirmayan guvenlik CLI araci.
2. smart-caliper: Alt piksel hassasiyetinde bilgisayarli goru ile parca olcumu ve uretim hata tespiti yapan denetim motoru.
3. QML: Qiskit platformu uzerinde Kuantum Makine Ogrenmesi ve klasik ogrenme modellerinin karsilastirmali basarim degerlendirmesi (Lisans Bitirme Tezi).
4. LLM-Comparison-Benchmarking: Buyuk dil modellerinin mimari kisitlara ve veri sozlesmelerine uyumunu olcen kontrollu benchmark calismasi.
5. 30DaysOfAI: 30 gun boyunca her gun bir yapay zeka uygulamasinin gelistirilip Streamlit ve Hugging Face uzerinde canliya alindigi maraton.

Iletisim:

• LinkedIn: suleyman-toklu10
• Kaggle: tiheli
• Hugging Face: tiheli
• Medium: @sulotkl
• E-posta: toklusuleyman24@gmail.com

---

_{Suleyman Toklu · 2026}