Engineer–Researcher building intelligent systems across data, learning, and orchestration.
I am a Python Engineer with a background in political science and economics research. I specialize in bridging the gap between advanced algorithms (Causal Inference, Optimization) and production engineering (Multi-Agent Systems, Open Source Tools).
My work focuses on building autonomous agents, designing real-time experimental platforms, and creating reliable open-source tools for the data science community.
I prioritize reproducibility, system architecture, and usability in my code.
Synthetic_Difference_in_Difference (Open Source Library)
Open Source Contribution: A production-ready Python library for the Synthetic Difference-in-Differences (SDID) estimator.
- Engineering: Designed, packaged, and published a library accessible via
pip install, ensuring standard distribution protocols. - Optimization: Implemented complex econometric algorithms with optimized matrix operations using NumPy and Pandas.
- Impact: Provides researchers with a stable, documented tool for causal inference and policy evaluation.
Carbon-Emissions-Trading-Platform (Real-Time System)
Complex System Design: A real-time, multi-player trading platform simulating carbon markets.
- Architecture: Built a real-time trading environment where multiple users interact simultaneously (bidding/asking) under varying constraints.
- Concurrency: Handled synchronized state updates and high-frequency data flow between the server and multiple clients.
- Stack: Built with oTree, WebSockets, JavaScript, and Python.
Social-Debate-AI (Multi-Agent System)
Agent Orchestration: A deep learning system where autonomous agents with distinct personalities debate to persuade a target audience.
- System Design: Engineered a modular Multi-Agent workflow using LangGraph, enabling complex state management.
- Algorithm: Integrates Graph Neural Networks (GNN) for social impact prediction and RL (PPO) for strategy optimization.
- Orchestration: Implemented dynamic agent interactions and memory handling to simulate realistic persuasion processes.
| Domain | Technologies |
|---|---|
| Languages | Python (Expert), SQL, JavaScript, R, C++ |
| AI & Agents | LangChain, LangGraph, OpenAI API, RAG, PyTorch (RL/GNN) |
| Data & Algo | Pandas, NumPy, Causal Inference (SDID), Optimization (GA/PSO/Bayesian) |
| Backend & DevOps | PyPI Packaging, FastAPI, Flask, Docker, GitHub Actions |
- FinFlash: Autonomous financial news analysis agent using Neural Search (Exa) and GPT-4o.
- Bert-Hyperopt-Comparison: Benchmarking genetic algorithms and Bayesian optimization for BERT fine-tuning.
- Algorithmic-Learning-Journal: A structured archive of data structures and algorithm design patterns.