Hi, I'm Mitash Shah ML Engineer & Quantitative Researcher

Graduate student at USC specializing in AI/ML and Quantitative Finance. Passionate about building intelligent systems and algorithmic trading strategies.

View My Work Get In Touch
AI/ML
Quant Finance
Software Dev
Research

About Me

Passionate about technology and innovation

My Journey

I'm a Master's student in Computer Science at USC with a strong foundation in both Artificial Intelligence and Quantitative Finance. My journey spans from developing cutting-edge NLP models to building algorithmic trading systems.

Current Research

Currently working as a Graduate Research Assistant at USC's vITAL Lab, I'm contributing to TruthTorchLM, an open-source library for evaluating LLM truthfulness. My research focuses on improving consistency assessment in AI-generated responses.

Technical Expertise

AI/ML

Python PyTorch TensorFlow NLP Deep Learning

Quantitative Finance

Algorithmic Trading Statistical Arbitrage Options Pricing Risk Management C++
USC Graduate Student
Master's in Computer Science with 3.8 GPA
Passions
Reading Dan Brown & Agatha Christie, Barcelona FC fan
Published Researcher
IEEE & EMNLP Papers
Open Source Contributor
TruthTorchLM Library with 25+ methods

Professional Experience

Building the future through research and innovation

Graduate Research Assistant

vITAL Lab, USC Sept 2024 - Present
  • Contributing to TruthTorchLM, an open-source library with 25+ truth assessment methods
  • Designed D-UE framework improving LLM consistency assessment by 30%
  • Leveraging ChatGPT, LLaMA, and DeBERTa for advanced NLP research

AI & Information Systems Intern

Fortive Jun 2024 - Aug 2024
  • Engineered LLM-based summarization tool improving data comprehension by 40%
  • Developed multilingual translation system with fine-tuned NLP models
  • Enhanced automated data visualization pipeline using Azure's distributed computing

Software Development Engineer

GM Infotech Jan 2023 - Nov 2023
  • Developed responsive web applications using React.js and Tailwind CSS
  • Migrated legacy sites to WordPress with SEO optimization
  • Integrated RESTful APIs using Flask and FastAPI

Featured Projects

Innovation across AI/ML and Quantitative Finance

Cross-Lingual Hate Speech Detection

Scalable multilingual hate speech detection system using XLM-RoBERTa with multi-task learning framework.

Python PyTorch Hugging Face CUDA
0.79 Macro F1 Score
15% Accuracy Boost

Algorithmic Trading Engine

Cointegration-based mean-reversion strategy across 3k liquid US equities with sub-10ms latency.

Python C++ IBKR API Bayesian Optimization
20% CAGR
1.24 Sharpe Ratio

Continual Learning Framework

Novel framework combining dynamic layer expansion with hybrid replay strategy for continual learning.

Python PyTorch Dynamic Architecture Hybrid Replay
50.41% Accuracy Improvement
2× Memory Efficiency

Monte-Carlo Options Pricer

Advanced pricing engine for exotic options with genetic algorithm calibration and stress testing.

Python QuantLib Genetic Algorithms Monte Carlo
±4% Market Accuracy
35% Runtime Reduction

Grid World Navigation System

Advanced pathfinding simulation using Q-Learning and Deep Q-Learning algorithms for dynamic environments.

Python TensorFlow Reinforcement Learning Q-Learning
40% Time Reduction
16x16 Grid Size

Publications

Contributing to the academic community

TruthTorchLM: A Comprehensive Library for Predicting Truthfulness in LLM Outputs

EMNLP 2025 (Accepted)

Open-source library featuring 25+ truth assessment methods for evaluating Large Language Model outputs.

Applications of AI and Blockchain in Healthcare

ISBN: 979-8-3503-1341-3

Comprehensive analysis of AI and blockchain technologies in healthcare applications.

IEEE Paper

Implementation of RFID-based Lab Inventory System

ISBN: 978-1-6654-9291-1

Innovative RFID-based system for efficient laboratory inventory management.

IEEE Paper

Get In Touch

Let's collaborate and build something amazing together

Email

mitashsh@usc.edu

Phone

(213) 756-8526

Location

Los Angeles, CA