Building intelligent infrastructure systems through low-level engineering and machine learning — from kernel-space debugging on enterprise WLAN stacks to transformer-based generative recommenders.
I'm a software engineer with a background in embedded Linux, enterprise networking, and low-level firmware. I previously worked at embedUR Systems on enterprise WLAN platforms — debugging kernel panics, memory leaks, multicast instability, VLAN edge cases, MTK driver integrations and SmartCast packet delivery on Ruckus enterprise access points running OpenWrt.
I'm now pursuing an MSc in Computer Science with an AI specialization at Leiden University, where my work spans transformer-based generative recommenders, reinforcement learning, and LLM evaluation for high-stakes domains.
The thread connecting all of it: building software that is fast, correct, and observable — whether it lives in kernel space or in a training loop.
A subset of work spanning generative recommenders, reinforcement learning, LLM evaluation, embedded vision and on-chain infrastructure.
Transformer-based generative recommender system using RQ-VAE semantic tokenization for sequential recommendation on Amazon review datasets.
Proximal Policy Optimization implementation with detailed analysis of training dynamics, reward convergence, and stability on classic control benchmarks.
Evaluation framework comparing GPT-4 with open-source LLMs on rewriting medical text into plain language while preserving clinical meaning.
AI-powered bill splitting application with receipt OCR, voice command parsing and direct bunq banking integration.
Blockchain-based incentive platform on the Solana ecosystem that tokenizes verified e-waste recycling actions.
Crowd-sourced incident reporting portal with on-chain verification on Avalanche / XDC.
Real-time accident-prevention system combining OpenCV-based hazard detection with Arduino sensor input.
Graduate research and coursework focused on generative recommenders, reinforcement learning and LLM evaluation methodologies.
Worked on enterprise access point firmware and OpenWrt-based platforms.
Graduated with 83.5% and awarded Best Outstanding Student of CSE — 2024.
A working notebook of debugging journeys, paper reproductions and systems deep-dives.
Reach out for research collaboration, engineering roles or technical conversation.