About Me
I'm a student at the University of Pennsylvania's Jerome Fisher Management and Technology Program, passionate about Machine Learning and Systems Engineering.
My goal is to bridge the gap between theoretical concepts and practical solutions, contributing to innovations that make a meaningful impact on how we interact with technology.
Experience
Machine Learning Engineering Intern
Apple | May 2024 - Present
- Reduced KPI prediction error by >10% using Mamba state-space model
- Deployed models on AWS, using Spark, Hadoop, and Airflow
- Automated Ad Review with fine-tuned Mistral LLM, increasing efficiency by 35%
Projects
Discus.ai
Open-source ML library for synthetic data generation and LLM fine-tuning.
- Gained 60+ stars and 300+ users in 2 months
- Implemented dual-phase LLM fine-tuning and retriever training
- Achieved performance increase of +8.9%
AutoRegressive KAN Transformer
Custom decoder transformer with Kolmogorov-Arnold Network.
- Developed from scratch in PyTorch
- Achieved log loss of 2.1 using 300,000 parameters
- Surpassed nanoGPT performance
UNIX Operating System
UNIX-like OS developed in C.
- Custom PCB and priority scheduler with0% starvation
- Used pthreads for concurrency
- Implemented custom FAT file system
Solar Lot Predictor
2nd Place at HackHarvard for Best Earth & Space Hack.
- Full-stack app to calculate solar farm sustainability
- Used Tensorflow CNN for satellite imagery analysis
- Achieved 96% accuracy in estimating solar farm costs
Technical Skills
Languages
Java, Python, C, C++, PostgreSQL, Javascript, GraphQL, CUDA, Triton
Frameworks & Tools
PyTorch, JAX, Flask, Typescript, Next.js, Kafka, Spark, Hadoop, Airflow, Tensorflow, JUnit, Git, Docker, Google Cloud, AWS, VectorDBs, MongoDB, DynamoDB, Kubernetes
Concepts
Machine Learning, Deep Learning, Distributed System Design, Cloud Computing, Object-Oriented Programming, Frontend, Backend, REST API & API Design, Agile Methodologies, Test-Driven Development