Kiarash Zamani.
Building ML systems for real-world perception and intelligent hardware. Currently developing a visual memory assistant on NVIDIA Jetson Orin Nano and contributing to satellite flight systems.
Professional Experience
Delta-Q Technologies
Embedded Software Engineer Intern
- →Architecting and validates low-level C firmware for high-voltage electronics, overseeing DC/DC control, power modes, and safety-critical fault handling.
- →Developing production-grade firmware for EV charging systems, meeting strict functional safety requirements for Tier 1 industrial OEMs.
- →Integrating automotive communication protocols including CANOpen, J1939, and J1772 to ensure seamless vehicle-to-charger networking.
- →Engineering an object-oriented Python/PySide6 automation tool that utilizes XML parsing to reduce hardware data transfer time by 60%.
- →Maintaining release momentum by triaging code backlogs, isolating root causes of firmware bugs, and coordinating fixes across branches.
ALEASAT — SFU Satellite
Full-Stack Software Engineer
Developed a real-time telemetry platform using React + Python to monitor and analyze spacecraft data streams.
Engineered a fault-tolerant data pipeline processing 200+ packets/sec with validation, retries, and data quality checks.
Optimized WebSocket streaming and batching to reduce dashboard latency by 30% under bursty workloads, and benchmarked telemetry against expected system models to identify reliability issues.
Technical Projects
Visual Defect Detector
Trained a YOLOv8 defect detector on a 1,000+ image dataset for aerial power line inspection, reaching 82%+ mAP with augmentation and contrast normalization.
Deployed an asynchronous FastAPI inference service with endpoint-level reporting for average / p95 latency and throughput.
Load tested 1–25 concurrent requests and optimized p95 latency with headless inference, sustaining about 5 req/s on CPU. Built a reproducible MLOps pipeline with automated preprocessing and artifact versioning to keep training and API outputs aligned.
Malware Signature Classifier
CNN-based pipeline using PyTorch to classify malware signatures as image-like bytemaps. Applied computer vision for binary classification with a Jupyter user interface.
MicroSentinel
Tamper-evident Arduino device using Embedded C++ to record physical access attempts.
SIMD Audio
8x speedup normalization pipeline using C++, LLVM, and the Parabix framework.
Vehicle Sensor Logger
Real-time C application for ESP32 with CRC checks for UART telemetry integrity.
Mountain Ride Share
Full-stack Node.js app using Google Maps API to track carbon emission savings.
Game Keeper
Android app for gameplay tracking. Built with JUnit 5 and Agile methodologies.
Anomaly Detection
Cybersecurity project using Hidden Markov Models and RL on power data.
EcoSpectra VR
Unity/C# VR experience for Oculus focusing on environmental protection.