Trustworthy AI2026
IBP Verifier
A from-scratch neural network verifier that computes sound output bounds for ONNX models under L∞ perturbations, using interval and symbolic bound propagation.
- Python
- NumPy
- ONNX
- pytest
SourceCase study
Projects
Things I've built
From certified robustness and verification tooling to drift monitoring, LLM applications, and from-scratch paper reproductions.
8 projects shown
- Trustworthy AI2026
α,β-CROWN Reproducibility Study
An independent reproduction of α,β-CROWN on the VNN-COMP CIFAR2020 benchmark: 147 robustness properties verified locally, with 93.2% agreement with the official verdicts.
- Python
- α,β-CROWN
- Pandas
- Matplotlib
- Jupyter
SourceCase study - MLOps2024 – 2025
Feature Drift Detection
A statistical monitoring pipeline that flags feature drift between reference and production data, with Kolmogorov–Smirnov tests implemented from scratch in NumPy.
- Python
- NumPy
- Pandas
- Scikit-learn
- Streamlit
SourceCase study - LLMs & RAG2025
MCP Server for PDF Q&A
A Model Context Protocol server that connects PDF documents to LLMs through an automated retrieval pipeline — extraction, chunking, embedding, and similarity search.
- Python
- MCP
- LangChain
- Embeddings
- Vector Search
SourceCase study - Computer Vision2024
Vision Transformer (ViT) Reproduction
From-scratch PyTorch reproduction of “An Image is Worth 16×16 Words” with a modular OOP design and custom training pipelines.
- Python
- PyTorch
SourceCase study - Signal Processing2026
Real-Time Acoustic Radar
Real-time motion detection from a laptop's own speaker and microphone: emit an inaudible 18–20 kHz tone and classify its Doppler-shifted reflections.
- Python
- NumPy
- SciPy
- Scikit-learn
- PyTorch
SourceCase study - Machine Learning2026
TinyGPT Language Model
A GPT-style language model trained on diverse text corpora (Amazon, IMDB, Reddit), covering architecture optimization and tokenization.
- Python
- PyTorch
- NLP
Case study - Cybersecurity2024
Network Intrusion Detection (NSL-KDD)
ML pipeline for network anomaly detection: EDA on ~125k records, Random Forest with Bayesian tuning, and an interactive Streamlit dashboard.
- Python
- Scikit-learn
- skopt
- Streamlit
- Pandas
Case study