Data scientist and ML engineer who ships real systems — from computer vision classifiers to NLP pipelines — and makes models do things that matter.
About Me
I'm a UC Berkeley Data Science grad (Dec 2025) with a focus on applied ML — the kind that lives in production, not just notebooks. I've deployed computer vision systems, NLP pipelines, and recommendation engines that drove measurable outcomes at real organizations.
My sweet spot is the intersection of modeling and engineering: I care about how a model is built, but I care just as much about how it gets integrated, monitored, and made useful for people who aren't data scientists.
Currently seeking full-time ML Engineer or Data Scientist roles where I can build systems at scale.
What I Work With
Tools and techniques I've used on real projects — not just tutorials.
scikit-learn, Random Forests, Gradient Boosting, A/B Testing, Feature Engineering
Image classification pipelines, TensorFlow, PyTorch, production deployment
Sentiment analysis, prompt engineering, generative AI integration, Claude API
Python, NumPy, pandas, SQL, automated ETL, daily delivery cadences
Matplotlib, Seaborn, audience segmentation, dynamic pricing, market analysis
Python, Java, SQL, C, Git, Jupyter, real-time model integration, API deployment
Career
Where I've built things and what came out of it.
Work
Side projects and competition work that pushed me further than internships did.
Resume market intelligence tool that maps live U.S. opportunities using local AI embeddings and semantic scoring. Features Signal Paths — three strategic next-move cards built from the shortlist.
Implemented slime mold network simulations in Python to optimize Berkeley transit routes. Integrated 3D city models with biological network algorithms to propose sustainable transport strategies.
Biodesign Challenge 2024Led AI-driven routing and demand prediction for a drone-based food delivery prototype. Used clustering algorithms and shortest-path heuristics with a built-in NLP assistant for stakeholder insights.
🏆 3rd Place — ENJAZ Tech CompetitionEnd-to-end NLP system integrating Claude API to analyze donor messages at scale, feeding a real-time outreach decision engine for a nonprofit platform.
Production — Helping HandsComputer vision pipeline deployed in a live nonprofit intake workflow. Reduced manual image review time by 40% through preprocessing optimization and batched inference.
40% time reduction in productionLet's Talk
Open to full-time ML Engineer and Data Scientist roles. If you're building something interesting, I'd love to hear about it.