Applied Data Science · Los Angeles, CA
Building intelligent systems at the intersection of machine learning, high-performance computing, and real-world impact
MS Applied Data Science student at USC with 3+ years building production ML, LLM, and analytics systems. I take problems from raw data to deployable models — RAG pipelines, agentic AI platforms, GPU-accelerated workloads, and decision analytics.
Currently seeking Fall 2026 co-op and New Grad 2026 roles in Data Science, ML Engineering, and Applied AI.
| Project | What it does | Stack |
|---|---|---|
| Agentic Trend Orchestrator | Agentic AI workspace for creators — trend discovery, team workflows, chat | FastAPI, Next.js, TypeScript |
| EcoMate-AI | Multimodal carbon footprint analyzer from receipts and daily activities | Python, FastAPI, Streamlit, GenAI |
| CUDA Python Library | GPU-accelerated matrix ops and image convolution benchmarks | CUDA, C++, Python |
| Quant Trading Backtester | End-to-end algo trading pipeline with XGBoost and backtesting | Python, XGBoost, pandas |
| BizScout Location Intelligence | Geo-intelligent restaurant site selection with heat maps | Python, Streamlit, geospatial |
| Green Food Purchasing Analytics | County-level sustainable food spending analysis (USDA data) | Python, Jupyter, NoSQL |
Agentic Trend Orchestrator — Full-stack agentic platform with auth, workflow boards, team assistant, and real-time chat. Live demo
EcoMate-AI — Won recognition at sustainability hackathons. OCR + emission factors + GPT-powered recommendations. Portfolio case study
CUDA-Accelerated Python Library — Custom CUDA kernels benchmarked against CPU baselines with reproducible artifacts and project report.
Languages Python SQL C++ TypeScript Bash
ML & AI PyTorch scikit-learn XGBoost Hugging Face LangChain RAG LLMs
Data pandas NumPy Spark Plotly Matplotlib Tableau
Engineering FastAPI Streamlit Docker CUDA AWS GCP Git
- Applying for Data Science & ML roles — Fall 2026 co-op / New Grad 2026
- Deepening LLMs, RAG pipelines, and agentic AI systems
- Building projects that connect real datasets to deployable models