Hello World (Yes, Really)
Why a Deep Learning Engineer who talks to machines all day decided to start talking to humans. Spoiler: the machines suggested it.
Experience
Deep Learning Engineer
RadiusAI Jun 2024 - Present- Built and maintained 10-15 production ML models for Fortune 10 clients (CVS) handling 5K different SKUs with 90% precision
- Deployed models using Triton inference server achieving sub-1 second inference latency
- Developed automated data quality pipeline reducing data corruption by 60%
- Integrated Vision-Language Models improving model accuracy by 3%
ML Developer
CARAIO Technologies (UC Berkeley Incubated) Jun 2022 - Aug 2024- Led 2-member team to develop end-to-end pipeline for medical chromosome analysis
- Achieved 85% accurate segmentations and 98% accurate classification with limited data
- Built complete prototype to initial product pipeline within 1 year (now rolled out to users)
- Part of founding team at UC Berkeley incubated startup
Technical Skills
ML/AI
Cloud & Infrastructure
Programming
Featured Projects
Chromosome Classification
Deep learning model for medical chromosome analysis achieving 98% accuracy
MWP-T5
Fine-tuned T5 model for generating solvable math word problems
Ground-up Series
GitHub repositories reproducing foundation model architectures from scratch
Open Source
Hugging Face Ecosystem
Contributing to transformers, datasets, and model implementations
PyTorch & Deep Learning
ArcFace Classification, Siamese Networks, and backpropagation implementations
Personal Projects
Hand detection, pose estimation, and from-scratch ML implementations
Pull Shark
Arctic Code
Quickdraw
Research & Publications
C3PO: A Lightweight Copying Mechanism for Translating Pseudocode to Code
ACLMWP-T5: Fine-tuning with Numeracy for Math Word Problem Generation
IEEE Xplore (To be published)Chromosome Segmentation Analysis Using Image Processing Techniques and Autoencoders
Research PaperLatest from WSYC Blog
Why Should You Care - Deep dives into tech topics that matter