My work sits between applied AI, backend systems, and connected infrastructure. I like building things from first principles, turning complex workflows into clear products, and working in early-stage environments where technical depth and execution speed both matter.
AI startup building a B2B marketplace and managed data operations layer for ML-ready datasets.
A data-centric AI venture focused on the supply side of machine learning: sourcing, validating, structuring, and commercializing proprietary datasets for AI teams.
Anthropic Opus 4.7 Hackathon 2026 build focused on automated labeling, confidence-based routing, and training-ready dataset quality.
A frontier data-operations system designed to compress one of the deepest bottlenecks in modern AI: turning raw data into cleaner, higher-confidence datasets through automated labeling, selective human review, and dataset-level evaluation.
LTE-connected IoT platform for real-time vehicle control, edge inference, and AI-ready data workflows.
A real four-machine connected mobility lab that combines LTE networking, distributed control, perception modules, Jetson-ready inference offload, and disciplined runtime validation under real hardware constraints.
SRTP vs SEPA Direct Debit — end-to-end flow simulation with real-time UI
Built to make SRTP tangible: a Flask API + WebSocket events + a live web UI that demonstrates the full Request-to-Pay lifecycle and the operational differences versus Direct Debit.
Project-based, peer-reviewed formation across C, Unix, concurrency, networking, and applied systems
Not a single repo or a “course artifact”—this is a multi-project engineering journey built through autonomous problem solving, peer evaluation, and sustained delivery across low-level and applied systems work.
Four-board embedded system combining custom hardware, LarVA-based RISC-V logic, sensing, GPS, and LoRa telemetry
A serious hardware-software integration build: four dedicated boards for power, compute, sensing, and wireless communication, tied together by a LarVA-based RISC-V subsystem, low-level firmware, and real telemetry workflows.
A small editorial layer for ideas worth expanding beyond a project card.
Apr 14, 2026
Prediction markets are becoming a compelling interface for AI agents: structured questions, live probabilities, public APIs, and a natural feedback loop for research and decision-making.
Mar 15, 2026
The most important shift in AI right now is not just better models. It is the emergence of protocols, tools, and evaluation layers that make agent systems operable.
