MILOSZ.EXE — Netscape Navigator 4.0

~ MILOSZ MISIEK ~

Software Engineer & PhD Candidate

HTML 3.2 Approved Powered by Coffee & Dial-Up
★ NEW! Welcome to my homepage — under construction since 1996 ★ Sign my guestbook! ★ Drop me a message below! ★
[Photo of Milosz Misiek]

About the Webmaster

Name: Milosz Misiek

Occupation: Software Engineer

VIEW WORK   CONTACT

EXPERIENCE

Period Type Role Company Focus Highlights Technologies
From: May 2023
To:   Present
 WORK Software Engineer Elastic Email Full Stack Development & AI Integration
  • Implementing AI-powered features using RAG architecture and large language models
  • Core engineer on Inbox — owning features across frontend, backend, and data layers
  • Building shared component library and reusable patterns adopted across product teams
  • Diagnosing and resolving performance bottlenecks across the full stack
 TypeScript, React.js, Node.js, Express.js, RAG, OpenAI API, MongoDB, pgvector, REST API, Git
From: Oct 2024
To:   Oct 2028
 ACADEMIC PhD Candidate West Pomeranian University of Technology Cybersecurity & Phishing Detection
  • Published 2 papers (HICSS 2025, IEEE Access 2026), 3rd accepted for CISIM 2026
  • Built browser extension for real-time phishing URL detection using XGBoost
  • Developing multi-layer detection pipeline combining NLP and computer vision
 Python, PyTorch, Deep Learning, NLP, Computer Vision, Cybersecurity
From: Mar 2023
To:   May 2023
 WORK Web Developer Maritime University of Szczecin Procurement System Development
  • Built and delivered the first version of a full-stack procurement system (NDA project)
  • Deployed and configured on Debian server with secure database management
  • Led development across the entire project lifecycle to on-time delivery
 React.js, Redux, Express.js, TypeScript, MySQL, HTML5, CSS3, Debian Linux
Software Engineer
May 2023 – Present
WORK
Elastic Email
Full Stack Development & AI Integration
  • Implementing AI-powered features using RAG architecture and large language models
  • Core engineer on Inbox — owning features across frontend, backend, and data layers
  • Building shared component library and reusable patterns adopted across product teams
  • Diagnosing and resolving performance bottlenecks across the full stack
TypeScript, React.js, Node.js, Express.js, RAG, OpenAI API, MongoDB, pgvector, REST API, Git
PhD Candidate
Oct 2024 – Oct 2028
ACADEMIC
West Pomeranian University of Technology
Cybersecurity & Phishing Detection
  • Published 2 papers (HICSS 2025, IEEE Access 2026), 3rd accepted for CISIM 2026
  • Built browser extension for real-time phishing URL detection using XGBoost
  • Developing multi-layer detection pipeline combining NLP and computer vision
Python, PyTorch, Deep Learning, NLP, Computer Vision, Cybersecurity
Web Developer
Mar 2023 – May 2023
WORK
Maritime University of Szczecin
Procurement System Development
  • Built and delivered the first version of a full-stack procurement system (NDA project)
  • Deployed and configured on Debian server with secure database management
  • Led development across the entire project lifecycle to on-time delivery
React.js, Redux, Express.js, TypeScript, MySQL, HTML5, CSS3, Debian Linux

RESEARCH

Year Title Abstract Authors Status Topics Link
2026 Phishing protection model based on machine learning vision detection approach NEW! Detects phishing pages from screenshots by fusing frozen SigLIP 2 and DINOv2 embeddings, with no fine-tuning. Evaluated on 70,785 web-page screenshots, the fused MLP reaches an F1 score of 0.9474 and an AUC of 0.9839. Miłosz Misiek; Tomasz Hyla Accepted Phishing Detection, Visual Classification, Deep Learning, SigLIP 2, DINOv2
doi:TBA
2026 XGBoost-Based URL Phishing Detection Method With Cross-Dataset Validation NEW! Conducted comprehensive cross-dataset validation of XGBoost for phishing URL detection across over 760,000 samples. The model achieved 90.7% accuracy, training 12.6x faster and using 3.6x less memory than comparable neural networks. Milosz Misiek; Tomasz Hyla Published Machine Learning, XGBoost, Phishing Detection, URL Analysis VIEW PAPER
doi:10.1109/access.2026.3672690
2025 Preventing Phishing Attacks with Browser-Based URL Detection Built and evaluated a machine-learning browser extension using XGBoost to detect phishing attacks in real-time. The solution accurately identifies malicious URLs while offering an exceptional balance of computational efficiency and interpretability. Miłosz Misiek; Tomasz Hyla Published Security, Phishing Detection, URL Analysis, Browser Extension VIEW PAPER
doi:10.24251/hicss.2025.874
Phishing protection model based on machine learning vision detection approach NEW!
2026
Detects phishing pages from screenshots by fusing frozen SigLIP 2 and DINOv2 embeddings, with no fine-tuning. Evaluated on 70,785 web-page screenshots, the fused MLP reaches an F1 score of 0.9474 and an AUC of 0.9839.
Miłosz Misiek; Tomasz Hyla
Accepted
Phishing Detection, Visual Classification, Deep Learning, SigLIP 2, DINOv2
doi:TBA
XGBoost-Based URL Phishing Detection Method With Cross-Dataset Validation NEW!
2026
Conducted comprehensive cross-dataset validation of XGBoost for phishing URL detection across over 760,000 samples. The model achieved 90.7% accuracy, training 12.6x faster and using 3.6x less memory than comparable neural networks.
Milosz Misiek; Tomasz Hyla
Published
Machine Learning, XGBoost, Phishing Detection, URL Analysis
VIEW PAPER
doi:10.1109/access.2026.3672690
Preventing Phishing Attacks with Browser-Based URL Detection
2025
Built and evaluated a machine-learning browser extension using XGBoost to detect phishing attacks in real-time. The solution accurately identifies malicious URLs while offering an exceptional balance of computational efficiency and interpretability.
Miłosz Misiek; Tomasz Hyla
Published
Security, Phishing Detection, URL Analysis, Browser Extension
VIEW PAPER
doi:10.24251/hicss.2025.874

WORK

Name Year Role Description Tech Image Link
Inbox: Communication Platform
[ NDA ] 		2025 ★ Core Engineer ★ Built a real-time communication system as a core contributor. Unified multiple messaging protocols into a single app, delivering a seamless cross-channel experience. Next.js, NodeJS, MongoDB, GraphQL, TypeScript, WebSocket [Screenshot of Inbox: Communication Platform] VIEW LIVE
Home Page: Elastic Email
[ NDA ] 		2024 Engineer Contributed to the Elastic Email marketing website. Focused on performance and rendering efficiency to improve user experience. Next.js, TypeScript, TailwindCSS [Screenshot of Home Page: Elastic Email] VIEW LIVE
Template Editor: Elastic Email
[ NDA ] 		2023 Engineer Worked on a complex, client-heavy email template editor, addressing drag-and-drop integration, state synchronization challenges, and rendering performance for complex layouts. React, TypeScript, Styled Components, WordPress Gutenberg [Screenshot of Template Editor: Elastic Email] VIEW LIVE
Inbox: Communication Platform
[ NDA ]
2025
★ Core Engineer ★
Built a real-time communication system as a core contributor. Unified multiple messaging protocols into a single app, delivering a seamless cross-channel experience.
Next.js, NodeJS, MongoDB, GraphQL, TypeScript, WebSocket
[Screenshot of Inbox: Communication Platform]
VIEW LIVE
Home Page: Elastic Email
[ NDA ]
2024
Engineer
Contributed to the Elastic Email marketing website. Focused on performance and rendering efficiency to improve user experience.
Next.js, TypeScript, TailwindCSS
[Screenshot of Home Page: Elastic Email]
VIEW LIVE
Template Editor: Elastic Email
[ NDA ]
2023
Engineer
Worked on a complex, client-heavy email template editor, addressing drag-and-drop integration, state synchronization challenges, and rendering performance for complex layouts.
React, TypeScript, Styled Components, WordPress Gutenberg
[Screenshot of Template Editor: Elastic Email]
VIEW LIVE

CONTACT

Let’s connect.

Have an idea or interesting problem? I’d love to hear about it.

DIRECT

contact@miloszmisiek.com

Links

Drop a line

Sign the Guestbook!