Machine learning intern · Costa Mesa, CA

Production demand forecasting for McDonald's, Taco Bell, Jack in the Box and Chipotle — thousands of locations, daily horizon. I work on the full model lifecycle end to end: 30+ engineered features drawn from weather, zip-code demographics, holidays, school calendars, national events and macroeconomic indicators; walk-forward cross-validation; hyperparameter tuning; back-testing against historical sales.

I cut forecast error from 12% to 5%. I led a team of interns on the data pipeline behind the model — scheduled crawlers behind a rotating proxy pool, a headless browser for pages that render client-side, and an ETL layer that lands everything in a cleaned, de-duplicated store. I designed the schema with range partitioning and a composite index so the feature joins stay fast.

I've run this full-time across two consecutive summers — CS 61C alongside it in 2025, CS 184 and CS 161 in 2026.

Residual Robotics · 4-person team · Berkeley

Robot policies don't generalize the way language models do, and a large part of the reason is data: the manipulation corpora that exist are small and narrow relative to what the problem needs. We're collecting mass UMI data from skilled laborers in India to widen that, to sell to robotics foundation model labs. The hardware is a handheld gripper you can carry into any environment and record real manipulation with, no robot required.

I own the hardware. I design the gripper in CAD — a full parametric assembly in Onshape, jaw linkage, camera mount, IMU tray, fiducial plate — and we design our own PCB around a Raspberry Pi Zero for the sensing stack. V2 came down from 636 g to 200 g through topology optimization, and the design-for-manufacturing work gets all 16 printed parts off the bed with zero slicer supports.

CS 184 final project · 2-person team with Ninad Atale · proposal site

Commercial FEM solvers make you re-solve from scratch for every new load case. For linear-elastic materials that is wasteful, because any load combination is a superposition of load cases you have already solved. We built a solver that generates its own training set by sweeping load cases automatically, trained a graph neural network on the mesh, and exposed the result as a viewer you can push on and watch deform in real time.

A mesher (Assimp to load CAD and COLLADA, a manifold check, TetGen to discretize into tetrahedra) feeds a validated FEM solver that passes nine correctness checks and generates the training set. On top sits a MeshGraphNets-style GNN in PyTorch Geometric, encode-process-decode over node and edge features. I sized its depth against a measured receptive-field bound rather than a paper default: a load moves the whole part, but L message-passing layers only reach L hops.

Scoped to linear elasticity and direct forces; no torques, no anisotropy. Prediction error is measured against our own solver's solution, not a commercial one, because matching topology across representations is its own problem.