Talia
Mekhayche

A wearable sleep apnea risk detection system tracking respiratory depth and rate in real time, streaming data to a mobile app and triggering automated alerts on anomaly detection. Hardware was conceived by reverse-engineering a retractable badge reel.

• Reverse engineered a badge reel to design a custom spring-loaded housing for a potentiometer, translating chest expansion into variable voltage signals.
• Read voltage digitally via SPI protocol and an MCP3008 ADC on Raspberry Pi 5 running QNX OS.
• Data pipeline via HTTP → Express/TypeScript REST API → PostgreSQL deployed on Railway.
• iOS app built in SwiftUI displays live breaths/min, depth classification, and warning alerts.

• No MacOS support for QNX required environment workarounds mid-hackathon.
• Missing kit hardware and 3D printing issues resolved on the fly during the event.
• Planned: machine learning layer to adapt thresholds to each individual's breathing patterns.

An autonomous robot assistant that bridges AI reasoning with physical manipulation. Voice-activated, visually aware, and capable of interacting with real-world objects — built as a complete hardware-software integration project.

• Body: Tracked rover chassis driven by Raspberry Pi 5 and L289N motor driver.
• Arm: Fully 3D-printed 6-axis robotic arm with Arduino servo control and inverse kinematics for precision positioning. Ultrasonic sensors for obstacle detection.
• Brain: Gemini 2.5 Flash for scene reasoning · Whisper ASR for voice input · ElevenLabs for speech output.
• Vision: Dual-camera system running YOLO11 for real-time object detection and spatial awareness.

• Conceived the concept, designed and 3D-printed all physical components, planned and soldered all circuitry.
• Diagnosed and resolved motor power distribution failure hours before the submission deadline.
• Wrote robotic arm firmware in C++ on Arduino, learning and implementing inverse kinematics within the hackathon window.

A functional miniature wind turbine designed and fabricated from scratch, converting rotational kinetic energy into real electrical output via a DC generator.

• Modeled an offshore turbine in Fusion 360, optimizing blade geometry for aerodynamic efficiency and 3D printability.
• Designed and printed a full nacelle and rotor assembly in PLA, housing a DC generator.
• Successfully converted rotational kinetic energy into electrical output — verified by illuminating an LED.

Engineered two complementary subsystems to address contaminated water: Column A treats, decontaminates, and allows flow — Column B fully impedes all contaminated water from passing through.

• Gravel + Fine Sand layers block large and small physical impurities.
• Biochar + GAC (Granular Activated Carbon) remove chemical contaminants.
• Resin layer reduces electrical conductivity by stripping dissolved salts (magnesium).
• Green Sand + Cotton Balls absorb and trap remaining micro-impurities before a wet mesh cloth outlet.

• Gravel dispersal layer distributes force of incoming water to protect clay barrier below.
• Clay layer acts as the primary water-impermeable barrier.
• 6× Sponges (60ml each) serve as an absorbent failsafe for any bypass seepage.
• Hydrophobic cloth at the base provides a secondary barrier when dry.

A wearable haptic glove interface that simulates physical resistance in VR, targeting surgical training simulations. Full-stack integration from custom embedded hardware through to a Unity game engine environment.

• Built on ESP32 to deliver real-time force feedback within VR environments.
• Developed a mechanical restraint system using MG90S servos to simulate physical resistance when gripping virtual objects.
• Integrated hardware with Unity (C#) for high-precision surgical training simulations.
• Designed and 3D-printed all ergonomic components in Fusion 360; managed all electrical wiring and signal routing for the full motor array.