|

Amazon Catalog System Services

MIT LIDS — Laboratory for Information and Decision Systems

UROP focused on configuration-based firmware rehosting with Penguin: reconstructed root filesystems from opaque vendor update images (unblob under fakeroot; bootable *.rootfs.tar.gz artifacts), validated boot under whole-system emulation for D-Link DNS-320 (BusyBox boot and interactive root shell), and investigated extractor design (fw2tar/unblob) for more reliable Linux rootfs selection — while building practical skills in Dockerized analysis environments, reverse engineering, and embedded systems debugging. Debugged rehosting failures using Penguin runtime artifacts (console.log, netbinds, mounts) and iteratively patched guest filesystems via config-driven static_files.

MIT — 6.106 Software Performance Engineering

Implemented malloc, realloc, and free using doubly-linked binned free lists, free-node splitting, coalescing, bitmaps, batched sbrk, compiler intrinsics, caching, and prefetching — optimized for latency and space utilization.

Amazon Business Marketing Tech

MIT — 6.110 Computer Language Engineering

Built a full compiler for Decaf (a C subset): lexical analysis, parsing, semantic analysis, SSA construction, optimization, and x86 codegen. Applied dead code elimination, constant propagation, and common subexpression elimination.

MIT Global Languages

Built a GPT-based language-learning app for Japanese practice — prompt design, conversational tutoring flow, and a simple full-stack UI around the model API. Explored how to best utilize LLMs to reinforce language learning techniques, and techniques to control the complexity level of the model's outputs.

MIT EECS

Helped students design and debug projects spanning audio/image processing, flood fill, graph search, Sokoban, Minesweeper, SAT solving, autocomplete, symbolic algebra, Scheme, and a LISP interpreter — testing, feedback, and office-hours support.

Amazon Fulfillment Tech

MIT WebLab

Fullstack@MIT

Conceptualized MapIT, a map-based platform for MIT students to find and share events, resources, and classes.

SWE@MIT

MIT

MacGregor House E Entry Chair— planning events, managing budgets, and supporting dorm community life.

MIT

Built a Python poker bot with Monte Carlo equity approximation (NumPy); experimented with tabular/deep Q-learning (PyTorch), ensembles, and distributed training under limited compute. Won $150.

Chantilly, VA

GPA: 4.67/4.0 · SAT: 1550. APs (scores 4–5): Computer Science A, Statistics, World History, English Language, Calculus BC, Chemistry, US History, Macroeconomics, Microeconomics, English Literature, US Government, Comparative Government, Psychology, Biology, Physics C: Mechanics. Additional coursework: Oracle 1 (SQL), Multivariable Calculus, Linear Algebra, Advanced Programming Dual Enrollment (Python).