Competition Projects

Kaggle Project
Chatbot Arena Human Preference Prediction

Chatbot Arena Human Preference Prediction

We trained Gemma-2-9B and Llama-3.1-8B with 4-bit QLoRA to predict human preferences, using AdamW and gradient accumulation to optimize log loss. To improve training efficiency and robustness, we tuned the context length and adopted length-sorted batching. We further improved performance by constructing a validation-weighted ensemble, yielding more stable and accurate preference probabilities.

Embeded Systems Design

Embeded System
Autonomous inventory management system

Autonumous Inventory Management System

Build an autonomous inventory management system that tracks items, expirations, and temperature in real time. Use an ATmega328p with a barcode scanner (UART), LCD, DS1631 sensor (I2C), and keypad, bridged to a Raspberry Pi over SPI. Integrate Firebase and a SwiftUI app for adding/removing items, viewing metadata, and receiving temperature alerts with instant two-way sync. Implement a compact state machine for add/remove/info workflows with debounced inputs. Deliverables: hardware schematics, microcontroller firmware, Pi client/server, and a complete mobile UI for autonomous, low-overhead inventory control.