🎓 Master's Student in Applied Data Science @ University of Chicago
🔍 Aspiring Data Scientist | Machine Learning Engineer | LLM Enthusiast
📍 Currently in Chicago | Fluent in Python, SQL, and cloud-based deployment
I am passionate about transforming data into strategic decisions and scalable systems. With cross-industry experience—from automotive forecasting to SaaS product analytics—I build end-to-end machine learning pipelines, deploy predictive models in production, and translate data insights into business outcomes.
Currently, I’m exploring Large Language Models (LLMs) and multimodal learning, integrating NLP pipelines with real-world applications such as recommendation systems and customer behavior modeling.
- 🔬 Research-driven with a strong foundation in deep learning, time series forecasting, and recommendation systems
- 🧠 Recently optimized a DNN-based credit default model
- 🌍 Interned at Tencent, Roland Berger, Deloitte, GM, and Continental Automotive across business analytics, risk modeling, and automation
- ☁️ Proficient in cloud deployment using AWS Lambda, Docker, and Google Cloud Platform
- 📈 Strong focus on interpretable AI, using SHAP/LIME to bridge models and decision-makers
- Fine-tuning open-source LLMs for Medical bilingual Q&A and enterprise knowledge retrieval
- Building LangChain-powered pipelines integrated with vector databases
- Creating an AI-driven movie recommendation engine using embedding and Wide & Deep architectures
- Contributing to GitHub repos on AutoML, time series modeling, and multi-task learning
-
📊 Customer Default Prediction (Verizon)
Custom DNN and XGBoost models with financial cost-based loss function
→ +9% business impact, deployed via Docker & AWS Lambda -
🎥 Netflix Movie Recommender System
Combined CF, deep learning(Wide&Deep), and GCP SQL backend for real-time recommendations -
📉 Cryptocurrency Quant Trading Strategy (MIT Collaboration)
Applied GARCH, LSTM, and momentum strategies using live data from Binance -
🧠 LLM Finetuning for Bilingual Medical Application (Ongoing)
Finetuning LLaMA-2 using LoRA and AlpacaEval for Chinese-English knowledge synthesis and applying contrastive learning and data augmentation for model distillation
Let’s build smarter systems and drive decisions through data & AI! 🚀