About
I'm an undergraduate at UC Berkeley studying Electrical Engineering & Computer Science. I'm interested in software engineering and backend systems.
I enjoy building systems that are technically sound but also helpful for people. Most of my projects involve some mix of full-stack development, data organization, and clean system design.
Currently, I'm a Software Engineering Intern at LG NOVA, where I work across a Flask/PostgreSQL backend and a React/TypeScript frontend. I previously interned at iPick.ai, where I built an LLM-based portfolio optimizer, a reinforcement learning pipeline of trading agents, and a graph-based platform that visualizes relationships between thousands of companies; at Maadhev LLC, where I built a custom ACH payment system and Shopify integrations; and at Claythis, where I worked on their AI-driven 3D model generation technology.
Away from the keyboard, I'm usually doing something active or yapping with my friends about nothing in particular. I've played soccer since I was 4 and still jump into pickup and intramural games whenever I can, and I've been a Barça fan for just as long, for better or worse. I lift regularly and I'm on a cut right now, which I track a little too publicly in the hobbies section. When I'm not doing any of that, I'm probably playing video games with the same friends, or planning the next trip; the most recent big one was summiting Kilimanjaro.
GitHub Activity
Experience
Software Engineering Intern
LG NOVA
Aug 2026 - Present
Santa Clara, CA
- Root-caused silent data loss in a platform-wide Flask/PostgreSQL endpoint capping all paginated downloads at 500 rows and shipped a keyset pagination redesign guaranteeing complete, deterministic results.
- Deployed the fix across Python and React/TypeScript repos, expanding semantic search coverage of PitchBook financial fields by 233%, backed by a 38-test suite verifying gap-free pagination across 1,200+ rows.
Machine Learning Engineering Intern
iPick.ai
Aug 2026 - Present
Berkeley, CA
- Built an LLM-based portfolio optimizer using the Claude API that segments holdings into investment tracks, prunes dominated positions, and provides per-stock and per-portfolio risk metrics.
- Engineered a reinforcement learning pipeline of three trading agents (power picks, hidden gems, and a short-target detector) that uses realized returns over a fixed holding window as the reward signal to update each agent's policy.
Software Engineering Intern
Maadhev LLC
May 2026 - Aug 2026
Cupertino, CA
- Built a custom ACH payment system (Python, Flask, Stripe API) integrated with a Shopify app, enabling bank transfer payments at 70% lower per-transaction cost than card processing.
- Engineered Stripe webhook handlers to sync payment and fulfillment data between Stripe and Shopify in real time, automating order confirmation across 50+ transactions.
- Rebuilt the Shopify storefront and checkout flow, consolidating a multi step checkout into a single page.
Software Engineering Intern
iPick.ai
Jan 2026 - May 2026
Berkeley, CA
- Architected a graph-based company relationship visualizer using React, D3.js, and PostgreSQL.
- Built a supplier/subsidiary relationship extraction pipeline across 5,000+ public company tickers, storing structured supplier/subsidiary edges in PostgreSQL to power graph-based visualization in the frontend.
- Integrated the Anthropic API to generate real-time news summaries, enriching company profiles with live information.
Software Engineering Intern
Claythis
May 2025 - Aug 2025
San Mateo, CA
- Created a Unity-based game in C# showcasing Claythis's AI-driven 3D model generation pipeline, implementing core weapon & health systems such as reloading, recoil, animations, armor, and healing.
- Tested and validated integration of the AI-generated 3D character models into Unity, verifying rigging, animation, and rendering behavior across the pipeline's outputs.
Projects
Falsetto
Detecting AI Generated Music
Python · PyTorch · Hugging Face Transformers · librosa
A deep learning system that detects AI-generated music using audio embeddings and attention-based classification through an end-to-end machine learning pipeline.
DEFCON
Quantifying Defensive Contributions
Python · PyTorch Geometric · scikit-learn · XGBoost · NumPy
A graph neural network-based soccer analytics platform that quantifies defensive contributions using player tracking and event data.
SoccerBoard
Forecasting the World Cup with machine learning.
Python · JavaScript · XGBoost · PyTorch · FastAPI · Docker
A soccer analytics platform that predicts match outcomes and expected goals using statistical models, machine learning, and interactive visualizations.
SaveBites
Fight food waste, one receipt at a time.
Python · JavaScript · Node.js · Express · MongoDB · React
A full-stack food waste reduction web app that scans grocery receipts using an OCR pipeline to automatically extract items, update pantry inventories, and generate recipe suggestions.
Wordle
A web-based Wordle clone with real-time feedback.
JavaScript · HTML · CSS
A browser-based Wordle game with game state management, guess validation, and real-time visual feedback. Features 3 different modes, and a clean, robust UI presented to 30+ peers.
Hexapod Robot
A 6-legged robot that senses and avoids obstacles.
C++
A hexapod robot with multi-directional movement, obstacle avoidance, and real-time sensor integration. I designed the breadboard layout and soldered connections for 18 servo motors, an ultrasonic sensor, and an LCD display.
Coursework
Hobbies
Spotify
Spotify
The best way I can describe my music taste is “I like what I like.”
Valorant
If I'm not playing sports, I'm probably playing video games with my friends. We play Valorant most often. We're lowkey terrible, but that's beside the point.
Current rank
Soccer & FC Barcelona
I've loved soccer since I was 4 and I played competitively until college. I was captain of both my club team and high school varsity team, made the all league first team all 4 years of high school, and finished with 45 goals and 20 assists over 4 years. You can still find me playing pickup or intramural games :)
I've been a Barça fan since age 4 because of my older brother. I catch most of their games, and honestly my mood heavily depends on the result. (jk… mostly.)
Latest result
Gym
I lift regularly and I'm currently on a cut. I log every weigh-in, and this graph updates as I do.
Since the start
Weight change (lb)
Adventures
Places pinned
46
trips and stops
Countries
6
so far
Continents
4
of 7
Highest summit
19,341 ft
Mt. Kilimanjaro
Upcoming
Patagonia
Chile · Argentina
Antarctica
Antarctic Peninsula
Past
Contact
I'm actively looking for opportunities. If you're a recruiter, engineer, or just want to chat about a project or opportunity, please don't hesitate to reach out!