Kunal Singh Bainsla
Experience
Pursuing a bachelor degree with a core focus on network defense, cryptographic systems, vulnerability research, and applying machine learning to automated threat detection.
- Specializing in system security and network vulnerability analysis
- Investigating security models for AI pipelines and adversarial robustness
- Collaborating on open source security and data science tooling
Designing and experimenting with deep neural network architectures, predictive classifiers, preprocessing pipelines, and recommendation algorithms.
- Built end-to-end loan prediction pipeline with rigorous evaluation metrics
- Implemented movie recommender systems with collaborative filtering
- Developed robust missing-data imputation and feature encoding routines
Experimenting with multi-agent orchestration, tool calling, local LLM evaluation, and securing agentic workflows against prompt injections.
- Testing autonomous agent workflows and decision loops
- Exploring defenses against LLM jailbreaks and adversarial inputs
Work & Projects
GitHub (6+)Library Management System
A modular Python-based management architecture to handle catalog records, member authentications, and borrowing operations with clean OOP principles.
ML Pipeline Loan Prediction
An end-to-end production ML pipeline for loan approval prediction covering outlier cleaning, feature engineering, model training, and hyperparameter tuning.
Movie Recommender System
A recommendation engine comparing collaborative filtering and content-based similarity models to generate personalized user movie suggestions.
Titanic Data Preprocessing Pipeline
Comprehensive data preprocessing architecture addressing missing value imputations, categorical encodings, and robust scaling on the Titanic dataset.
Heart Disease & Bulldozer Price Prediction
Predictive machine learning models built alongside Andrew Ng's curriculum, featuring classification for cardiac risk and regression for heavy equipment valuation.
Data Science & ML Experiments
Curated repository of algorithmic experiments, statistical data investigations, and foundational deep learning implementations.
Blog
3 articlesA deep dive into Adversarial Machine Learning & AI Security
How imperceptible noise perturbations manipulate neural network embeddings and why defensive distillation is critical for production AI systems.
Building Production-Ready ML Pipelines with Scikit-Learn & PyTorch
A systematic guide to structuring leak-free data preprocessing, reproducible feature transformers, and automated validation gates.
Modern Recommender Systems: Collaborative vs Content-Based Filtering
Understanding matrix factorization, cosine similarity metrics, and cold-start mitigations in modern recommendation engines.
Skills & Technologies
Get in touch
Whether you are looking for an intern in Cyber Security / ML, want to collaborate on open-source projects, or just want to chat about AI security, feel free to reach out.