Margi Pandya

MS Student @ UC San Diego | Applied ML · Multimodal Learning · Vision & Language

Hi! I’m a Master’s student in Electrical and Computer Engineering at UC San Diego, specializing in machine learning for computer vision, language, and multimodal systems. I focus on building intelligent, scalable AI models that bridge research and real-world deployment.

What I Build

• End-to-end training and fine-tuning pipelines for vision-language and multimodal models using PyTorch and JAX
• Post-training methods including supervised fine-tuning and RL-based alignment to control model behavior
• Scalable ML systems with mixed-precision training, distributed data parallelism, checkpointing, and evaluation

Technical Skills

• Core Focus: Multimodal LLMs, Vision-Language Models, Post-Training & Alignment, Reinforcement Learning
• ML Frameworks: PyTorch (expert), JAX, HuggingFace, TensorFlow, Scikit-Learn
• Systems & Scaling: Distributed Training (DDP/FSDP), Mixed Precision, CUDA, Docker, Linux
• DevOps & Cloud: Git, CI/CD, MLOps, Azure, GCP
• Languages: Python, SQL, C/C++, Bash

Currently working on: multimodal post-training, RL-based alignment techniques, and scalable training pipelines for vision-language models.

I’m seeking full-time Applied ML / ML Engineer roles focused on post-training and alignment of foundation models, where I can apply fine-tuning, RL-based optimization, and scalable AI deployment.