Context
I build backend systems where design decisions surface as latency, failure modes, and cost.
My focus is reducing complexity — not by hiding it, but by structuring systems so their behavior remains predictable in production.
Experience
Infrastructure Engineering Senior Advisor
Building backend systems with a focus on failure behavior, observability, and correctness under production load.
Software Engineer
Built backend systems and internal tools for telecom workflows, improving behavior under high-throughput production traffic.
Graduate Software Developer
Shipped backend systems in production while strengthening monitoring, debugging workflows, and deployment reliability across multi-service environments.
Principles
Systems should fail predictably in production, not silently.
Hidden complexity is worse than visible complexity.
Good software makes operational behavior explicit.
What I Focus On
Designing backend systems that behave predictably under load.
Modeling failure modes and making them observable.
Reducing complexity by making system behavior explicit.
Selected Work
System Design Simulator
Simulation · ArchitectureDesigned and built an interactive platform for simulating backend architectures in real time — making architectural tradeoffs executable.
Stack: Flutter Web, Rust, WebAssembly
Impact: Reduced architecture iteration loops by ~35% in prototype reviews, enabling real-time tradeoff simulation across latency and cost.
Technical hook: Built around an event-driven simulation engine with async state propagation across service graphs.
Tradeoff: prioritized event-driven updates over polling to reduce latency spikes under load.
LuggageBot
Robotics · Computer VisionBuilt an autonomous robotics system that tracks and follows a human using real-time computer vision.
Stack: Python, ROS, OpenCV
Impact: Improved target-following stability by ~30% in noisy test runs while maintaining real-time tracking and control responsiveness.
Technical hook: Optimized a real-time vision pipeline to reduce frame-to-control latency and stabilize tracking under noisy motion.
Tradeoff: prioritized deterministic control loops over aggressive path optimization to keep behavior stable in noisy conditions.
Poker Planning
Collaboration · Real-time UXBuilt a planning poker tool that makes estimation sessions less painful and slightly more fun.
Stack: React, Django, WebSockets, Rive
Impact: Enabled instant session creation with synchronized voting and reveal, reducing coordination overhead in team estimation.
Technical hook: Implemented real-time vote synchronization with coordinated reveal logic to ensure consistency across participants.
Interaction layer: Integrated Rive animations to provide responsive visual feedback during user actions (validation, voting, session flow).
Tradeoff: prioritized responsiveness and simplicity over deep session features to keep live interactions smooth and distraction-free.
Pixel Pals
Systems · Algorithms · Game SimulationBuilt a console-based game where player and enemy behavior emerge from grid state and pathfinding logic.
Stack: C++, Console I/O
Impact: Implemented deterministic enemy pursuit using BFS, ensuring consistent behavior across runs while maintaining real-time responsiveness.
Technical hook: Modeled the game world as a 2D state grid and used breadth-first search to compute shortest paths for enemy movement.
Tradeoff: prioritized deterministic pathfinding over unpredictability to make system behavior easier to reason about and debug.