I'm a curious, high-ownership builder who turns ambiguous problems into something real. I learn by building, testing, questioning assumptions, and iterating, not by waiting for a perfect roadmap. I'm comfortable stepping into unfamiliar areas, working with different people, and taking an idea from "What if?" to a working solution.

With a background in Electronics & Telecommunication and an M.S. in Data Science (GWU), I work across agentic AI, applied machine learning, and evaluation: from safety layers for physical AI to turning unstructured information into trustworthy, usable knowledge. I care about both the technical details and the people affected by what I build.

Curiosity, persistence, humility, and a bias toward action are what I bring, and I'm always looking for the next hard problem worth solving.

• 🌍 Former IMF Data Science Extern, Washington, DC
• 🏛️ Delegate: World Bank Group Youth Summit ('25, '26) · International Youth Conference ('26) · GLF Nairobi ('22)
• 📊 Research on Evaluating Causal, Temporal, Multi-Hop, and Counterfactual Reasoning in Financial Large Language Models

skills = {
    "languages": ["Python", "TypeScript", "SQL", "R"],
    "ai_ml":     ["PyTorch", "Hugging Face", "Transformers", "scikit-learn"],
    "agentic":   ["LangGraph", "LangChain", "RAG", "LLM Evaluation"],
    "backend":   ["FastAPI", "Node.js", "PostgreSQL", "Supabase"],
    "tools":     ["Docker", "Git", "Streamlit", "Power BI"],
    "interests": ["AI Alignment", "Reliable AI Systems", "Human-Centered Design","Reinforcement Learning"],
}

Mechanistic Interpretability · AI Alignment · Agentic Evaluation · Reliable AI Systems · Reinforcement Learning

• AI Models Waste Up to 20% Compute: Thinking About Nothing
• When Gradient Descent Runs Out of Tangents
• Mathematical Optimization