AI & ML interests

None defined yet.

Recent Activity

dineth554Ā  updated a Space 2 days ago
Pnny13/legion-hurtyl
dineth554Ā  published a Space 2 days ago
Pnny13/legion-hurtyl
dineth554Ā  updated a Space 3 days ago
Pnny13/README
View all activity

Organization Card
Typing SVG

Visitors Status Team Focus


šŸ’€ Who We Are

Deathly Legion is a hands-on cyber team. We train AI models, dig into machine learning research, and run security operations. Not a brand — a team that actually builds things.

  • We fine-tune models, not just run them
  • We do real security research, not just CTFs
  • We write the tools we use
  • Small team. High output.


šŸ¤– AI & Machine Learning

What we actually do:

  • Build and fine-tune neural networks from scratch
  • Clean and prep datasets — the unglamorous work that makes models work
  • Optimize models for speed and size without trashing accuracy
  • Evaluate outputs properly, not just vibe-check them

Frameworks we use daily

           

PyTorch TensorFlow Keras scikit-learn HuggingFace NumPy Pandas


šŸ›”ļø Cybersecurity

Security research, threat modeling, system hardening, network analysis. We're not red team cosplay — we find actual problems and write actual reports.

Tools we know well:

Kali Wireshark Metasploit Burp Suite Nmap OWASP



šŸ’» Languages & Dev Stack

               

Python JavaScript Bash SQL Docker Git Linux Jupyter VS Code Google Colab Kaggle


šŸ“ˆ Training Focus Areas

Area What We're Working On
šŸ‘ļø Computer Vision Object detection, segmentation, custom classifiers
šŸ—£ļø NLP Fine-tuning LLMs, tokenizer work, RAG pipelines
šŸ“‰ Time Series Anomaly detection, forecasting models
🧠 Classification Multi-label, imbalanced datasets, edge cases
šŸ” Anomaly Detection Unsupervised methods for security applications
šŸ—‚ļø Dataset Work Scraping, cleaning, labeling, augmentation

šŸ† Team Structure

šŸ’€ Deathly Legion
ā”œā”€ā”€ šŸ¤– AI/ML Engineers      → model training, fine-tuning, deployment
ā”œā”€ā”€ šŸ” Security Researchers → vuln research, tooling, threat analysis
ā”œā”€ā”€ šŸ“Š Data Scientists      → dataset pipelines, analysis, evaluation
└── āš™ļø  Infrastructure      → compute, automation, environment setup

āš–ļø Ethics

We build responsibly. That means:

  • Models don't get deployed without evaluation
  • Security research stays legal and disclosed properly
  • No data gets used without checking its provenance
  • We don't cut corners on privacy

Footer typing

Made with Powered by