Skip to content
View ulpati's full-sized avatar
๐ŸŽฏ
Focusing
๐ŸŽฏ
Focusing

Block or report ulpati

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Content in all repositories owned by your account will be closed.
Maximum 250 characters. Please donโ€™t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this userโ€™s behavior. Learn more about reporting abuse.

Report abuse
ulpati/README.md

๐Ÿ‘‹ Hi, I'm ulpati

Currently working in Application Security. Passionate about cybersecurity, machine learning, automation, and building tools that solve real-world problems.


๐Ÿ’ผ Current Role

Application Security Associate

  • ๐Ÿ”’ Application security and vulnerability management
  • ๐Ÿ›ก๏ธ Collaboration with development teams on secure software practices
  • ๐Ÿ” Security assessment and risk analysis
  • ๐Ÿ“Š Working with security frameworks and industry standards

๐ŸŒฑ Currently Learning

  • ๐Ÿ” Application Security - Secure software development and vulnerability assessment
  • ๐Ÿ› ๏ธ DevSecOps - Integrating security into development workflows
  • ๐Ÿ“‹ Security Frameworks - Industry standards and best practices

๐Ÿš€ Projects

Machine learning system that predicts ATP match outcomes, evaluated on matches it never saw.

Performance: 66.3% accuracy, 0.603 log loss and 0.731 AUC on 19,086 held-out matches from May 2019 to May 2026, ahead of Elo-only (64.3%) and ranking-only (63.9%) baselines. The model used for predictions is then retrained on all 190,620 matches since 1970.

Key Features:

  • XGBoost with a strict chronological train, validation and test split
  • Overall and surface Elo ratings with experience-based K-factor
  • 80 leakage-free features: form, serve and return stats, workload, head-to-head, ranking
  • Tests proving that features never depend on the match result and match between training and prediction
  • Model explanations with TreeSHAP
  • Pickle-free model storage verified with SHA-256 manifests; hash-checked data download
  • Reproducible from a clean clone: atp download and atp train recreate every reported number; optional CUDA training
  • CLI with 11 commands, 170 tests (97% coverage), CI with CodeQL, Bandit and pip-audit

Tech Stack: Python, XGBoost, Scikit-learn, Pandas, NumPy, GitHub Actions


๐Ÿ“š Repository Collection

Bash tools for privacy, publishing and productivity, built with a secure development life cycle.

Tools:

  • ๐Ÿ›ก๏ธ Metadata Cleaner: removes EXIF, GPS, document and notebook metadata, redacts secrets and personal data in text files, and audits the result.
  • ๐Ÿš€ Publish Repo: initialises a repository, scans pending files for secrets, commits after confirmation and pushes without touching existing branches.
  • ๐ŸŽฎ Life Game: gamified habit tracker with levels, badges, a Pomodoro timer and Markdown study plan import.

Security: STRIDE threat model, no eval or sourcing of data, helper programs that never write next to the user's files, 87 bats tests including injection and symbolic-link cases, CI with ShellCheck on every change.

Tech Stack: Bash, ExifTool, qpdf, mat2, Git, GitHub CLI, ShellCheck, bats


Study materials, notes, and projects from certifications and online courses.

Completed:

  • ISC2 Certified in Cybersecurity (CC) (2026): security principles, incident response, access control, network security, security operations
  • CCNA 200-301 (2025): Networking fundamentals, OSPF, VLANs, IPv4/IPv6, ACLs, NAT, QoS, network security, automation
  • FreeCodeCamp - Data Analysis with Python (2024): Pandas, visualization, statistical analysis
  • FreeCodeCamp - Scientific Computing with Python (2024): OOP, algorithms, testing

Tech Stack: Python, Networking protocols, Cisco IOS


End-to-end data science projects demonstrating the complete ML workflow (2023-2024).

Projects:

  • Machine Learning: Wine classification with Random Forest (98% test accuracy)
  • Data Visualization: Global food production environmental impact analysis
  • SQL Analysis: U.S. travel safety analysis with PostgreSQL
  • Python Automation: Intelligent file organization system
  • Final Project: Student performance prediction with statistical modeling

Skills: Data wrangling, EDA, feature engineering, model evaluation, interactive visualization

Tech Stack: Python, Scikit-learn, Pandas, NumPy, Matplotlib, Seaborn, Plotly, PostgreSQL


Computational physics simulations and scientific computing projects from university coursework (2020-2022).

Statistical Physics (2022)

  • Monte Carlo simulations (Metropolis & Wolff algorithms)
  • 2D Ising model: phase transitions and critical phenomena
  • Random walks: diffusion and self-avoiding walks
  • SIR epidemic model on complex networks

Experimental Data Processing Lab (2021)

  • Numerical methods: integration, differential equations, root finding
  • Object-oriented design with C++
  • ROOT framework for scientific visualization

Computer Science with C++ (2020)

  • Fundamentals: algorithms, data structures, pointers
  • Structured programming and modular design

Tech Stack: C++, Python, ROOT, NumPy, Matplotlib, Jupyter


๐Ÿ› ๏ธ Technical Skills

Languages & Frameworks

python cplusplus bash

Data Science & Machine Learning

xgboost scikit-learn pandas numpy matplotlib seaborn jupyter

Tools & Platforms

linux git postgresql

Specializations

  • Application Security: secure development life cycle, STRIDE threat modelling, secure code review, supply-chain hardening (pinned actions, hash-locked dependencies, CodeQL)
  • Information Security: Network security fundamentals, security best practices, continuous learning in cybersecurity
  • Networking: TCP/IP, routing protocols (OSPF), VLANs, network configuration and troubleshooting (CCNA level)
  • Automation & Scripting: Bash scripting, Python automation, file processing, workflow automation, privacy tools
  • Machine Learning: Classification, regression, ensemble methods, leakage-free feature engineering, time-based validation, model interpretability
  • Data Science: Feature engineering, EDA, statistical analysis, data visualization, predictive modeling
  • Scientific Computing: Monte Carlo simulations, numerical methods, computational physics, ROOT framework

๐Ÿ“Š GitHub Stats

Top Languages

GitHub Stats


Securing systems โ€ข Building tools โ€ข Sharing knowledge through code

Pinned Loading

  1. atp-tennis-predictor atp-tennis-predictor Public

    Predicts ATP tennis matches with XGBoost on leakage-free features: overall and surface Elo, ranking, form, serve and return statistics, head-to-head. 66.3% accuracy on 19,086 matches the model neveโ€ฆ

    Python

  2. bash-scripts bash-scripts Public

    Three security-hardened Bash tools for Linux: a metadata cleaner for photos, videos, PDF, office files and notebooks; a Git publisher that blocks secrets before they are committed; a terminal habitโ€ฆ

    Shell

  3. data-science-portfolio data-science-portfolio Public

    Projects from the start2impact Data Science Master: PostgreSQL analysis of travel safety, Plotly study of food production, population and climate change, a Python file organizer, wine classificatioโ€ฆ

    Jupyter Notebook

  4. academic-physics-projects academic-physics-projects Public

    University computational physics projects (2020-2022): Monte Carlo simulations of the Ising model, random walks and SIR epidemics on networks in Python; numerical methods, data analysis with ROOT aโ€ฆ

    Jupyter Notebook

  5. certifications certifications Public

    ISC2 Certified in Cybersecurity (CC); study notes and projects from Cisco CCNA 200-301 (routing, switching, security, automation) and freeCodeCamp Scientific Computing and Data Analysis with Pythonโ€ฆ

    Python