Iโm a Data Scientist specialized in Artificial Intelligence (AI) and Cybersecurity. I build production-ready AI systems that are secure, privacy-preserving, and designed for real-world impact across multiple industries. Iโve worked with public-sector authorities on GDPR-compliant AI, privacy-preserving techniques, and secure handling of sensitive data, combining hands-on engineering with regulatory and risk awareness.
A core part of my work is failure-mode thinking: understanding how AI systems break through data leakage, weak access controls, insecure integrations, or poorly governed pipelines, and designing architectures that reduce those risks.
Machine Learning (11) | Cybersecurity (12) | Other (12)
Latest
Educational project demonstrating image-based malware family classification: classifying malware from pre-rendered grayscale byteplots, training a CNN from scratch, fine-tuning pretrained architectures (ResNet-50, ResNeXt-50, Inception-v3), and evaluating them.
Features:
- 25 malware families - 9,339 pre-rendered grayscale byteplot images.
- How malware can be represented as images โ what byteplot layout and texture encode.
- How convolutional networks learn family-specific structure from byte sequences.
Machine Learning (11)
A curated collection of machine learning projects covering various domains including computer vision, predictive modeling, cybersecurity, and data analysis.
| Repository | Description | Language |
|---|---|---|
| malware-classification | Image-based malware family classification from byteplots with CNN from scratch and transfer learning | Jupyter Notebook |
| spam-detection-using-AI | End-to-end SMS spam classification with scikit-learn and a Streamlit web app | Jupyter Notebook |
| E-commerce-Conversion-Prediction | Predicts whether an e-commerce visitor session will result in a purchase | Jupyter Notebook |
| Convolutional-Neural-Networks-CNNs-from-scratch | CNNs from scratch inspired by EfficientNet and Inception, trained on CIFAR-10 | Jupyter Notebook |
| CAN-T-SEE-THE-FOREST-FOR-ALL-THE-TREES | Univariate analysis with predictive modeling using decision trees and random forests | Jupyter Notebook |
| fine-tuning-an-local-llm-for-web-scraping | Fine-tunes Qwen2.5-3B to extract structured JSON data from job postings | Jupyter Notebook |
| digit-classifier-api | MNIST classifier served through a FastAPI REST API โ train, deploy, consume | Jupyter Notebook |
| digiti-signum | Classical and forensic fingerprint matching techniques on the SOCOFing dataset | Jupyter Notebook |
| digit-recognition | Streamlit app recognizing handwritten digits 0-9 with a CNN trained on MNIST | Python |
| Understanding-Data-Augmentation | Measures the impact of data augmentation on ResNet18 training on CIFAR-10 | Jupyter Notebook |
| data-processing | Chicago crime data 2008-2012 analyzed against census data for socioeconomic correlations | Jupyter Notebook |
Cybersecurity (12)
A curated collection of 12 cybersecurity projects spanning operational security, defense evasion, steganography, network security, system hardening, and hardware-based EDR evasion.
| Repository | Description | Language |
|---|---|---|
| project-horseshoe | Subnet scanner for local network discovery and TCP port enumeration | Python |
| wifi-penetration-testing | WiFi pentest guide: handshake capture, cracking, rogue AP detection, automated Python tool | Python |
| defense-evasion | MITRE ATT&CK defense evasion reference with filesystem manipulation and detection strategies | Python |
| project-bytehide | Byte-level file techniques: JPEG data hiding, polyglot files, codec-aware MP3 embedding | Jupyter Notebook |
| canary-tokens | Canary token internals: how they work, how to deploy them, how to detect them | Shell |
| Amnesiac-Shell | On-demand private shells with automatic session history erasure for OpSec | Shell |
| Changeling | MAC address randomization via NetworkManager to reduce WiFi and Ethernet tracking | Shell |
| Ikigai | Cross-distro Linux hardening utility | Shell |
| one-persons-feature-is-another-persons-exploit | Image metadata leakage risks for OPSEC and how to inspect and remove metadata | Jupyter Notebook |
| playwright-pixelscan | Inspects what websites can collect from a web browser using Playwright mobile emulation | Python |
| mouse-jiggler | Cross-platform Bezier curve hardware mouse jiggler on RP2350-One spoofing a USB mouse | Python |
| image-to-code | Embeds images into Markdown as Base64 โ self-contained, no external hosting | Python |
Other (12)
| Repository | Description | Language |
|---|---|---|
| clone-sweeper | Automated GitHub repository analytics collecting clone traffic data | TypeScript |
| pages | Blog and portfolio site with terminal-inspired design and auto-fetched GitHub projects | TypeScript |
| todays-command | Linux desktop widget delivering a daily command with quizzes and streak tracking | TypeScript |
| myjam | Audio converter built on ffmpeg with CLI and web GUI interfaces | TypeScript |
| podcast-retriever | Aggregates podcast data from iTunes Search API, RSS feeds, and Podchaser | Python |
| dynamic-webscraper | Extracts and structures data from any website without hardcoded parameters | Python |
| media-compressor | Compresses images and videos to target file sizes โ Flask, Tailwind, Docker | HTML |
| Atlas | Web-scraping toolkit for Swedish authorities: pagination, PDF crawling, metadata extraction | Python |
| gif-maker | Web tool generating GIFs from images or video โ React and Vite frontend, Flask backend | JavaScript |
| ai-job-finder | Multi-agent system automating job search and career guidance with CrewAI and LLMs | Python |
| ai-agent | AI agent system combining LLMs with custom tools and memory retention | Python |
| api-forge | AI-powered OpenAPI specification generator | TypeScript |
-
Project Foresight
Try it out โ
Note: since this is a WIP, it may not be accessible at all times.
Predicts weather temperature using machine learning and automated workflows.
Features:
- Integrates real-time weather and traffic data.
- Outputs weather temperature predictions based on real-time data.
- Automates data ingestion, processing, and ML predictions.
