About

I am an early-career engineer with a background in cybersecurity and a long-term goal to work on:

1. AI infrastructure / ML systems (primary), and
2. Quantitative / trading-style systems (secondary).

Rather than chasing buzzwords, I am deliberately building:

• Strong fundamentals in Python, C++, algorithms, and distributed systems
• Practical ML skills in PyTorch and MLOps
• A security-informed mindset for designing robust, failure-aware systems

a

My current focus

1. AI Infrastructure and ML Systems

I am following a structured roadmap centered on:

• PyTorch first: tensors, autograd, CNNs, training loops
• Distributed systems: Designing Data-Intensive Applications (DDIA) and MIT 6.824
• ML system design: reading Chip Huyen and studying real-world ML platforms
• MLOps tools: MLflow, Docker, Ray, Kubernetes basics

The goal is to be able to:

• Design and explain an end-to-end ML system (data → training → serving → monitoring)
• Discuss trade-offs in model deployment, scaling, and reliability
• Implement and debug real systems, not just notebooks

2. Security-informed Engineering

Coming from cybersecurity, I care about:

• How systems fail under adversarial or unexpected inputs
• How to reason about observability, logging, and forensics
• Reducing attack surface and building defense-in-depth into ML systems

This perspective is especially relevant for:

• Fraud detection (e.g., payments, abuse detection)
• Security AI (LLMs + security pipelines, anomaly detection)
• Mission-critical ML where reliability matters as much as accuracy

3. Quant-leaning Backup Path

As a secondary trajectory, I am exploring:

• Low-latency systems & networking
• Probability / statistics / expected value problems
• Architectures used by trading firms and exchanges (market data, order matching, risk systems)

This is not a hobby; it is a deliberate hedge in case the best-fit opportunities arise in quant / trading infrastructure rather than “traditional” tech.

How I work

• Plan-driven: I follow a written, time-bounded roadmap instead of jumping randomly between topics.
• Project-first: I anchor learning in concrete projects (e.g., ThreatLens, static site generator) rather than pure theory.
• Feedback-seeking: I iterate based on feedback from instructors, peers, and real interview-style problems.

What I want next

I am looking for opportunities where I can:

• Work on training, serving, or data platforms for ML
• Contribute to systems that require performance + reliability
• Use my security background to improve robustness and safety of ML-powered products

If that sounds like your team, please see my Projects and reach out.