ML Engineer • Software Engineer
I Just Try
I am an engineer dedicated to solving open-ended problems where software systems meet machine learning. My background in competitive programming provides the algorithmic foundation for my current work: building intelligent systems that are statistically rigorous, privacy-compliant, and aligned with user expectations.
🛡️ Synth Studio | Repository • Site
Privacy-first synthetic data generation platform for regulated industries.
- Privacy Engineering: Implemented Differential Privacy (DP) using RDP accounting for mathematical safety guarantees.
- ML Synthesis: Integrated CTGAN, TVAE, and Gaussian Copula models for high-fidelity tabular data.
- Automation: Built automated HIPAA/GDPR compliance pipelines based on re-identification risk metrics.
An Argument Navigator for scientific research that maps intellectual debates.
- Concept: Transitions from citation-based graphs to argument graphs (Builds Upon, Refutes, Extends).
🧠 The Conversational AI Tax | Research
A deep dive into failure modes within mental health AI systems.
- Methodology: Processed 30k+ reviews using BERTopic and RoBERTa for semantic intent analysis.
- Key Discovery: Identified a 300% spike in user distress complaints linked specifically to AI "memory" and "personality" failures following model updates.
- 🏛️ MIT Emerging Talent Program: Advanced coursework in Computer and Data Science, specialized in real-world data engineering.
- 🌱 Current Focus: Deep Learning architectures, Large Language Models (LLMs), and Scalable Data Science.
|
|
|
"Always a student of the craft."