Hello, I am Yugal
I am
Currently based in Los Angeles, CA, and pursuing a Master of Science in Computer Science at the University of Southern California, I am passionate about software development and creating impactful, innovative technological solutions. With a strong foundation in machine learning, deep learning, and full-stack development, I am constantly learning, coding, and striving to push the boundaries of technology to build meaningful experiences.
About Me
I am a passionate coder who thrives on transforming ideas into reality through software. With a deep commitment to innovation, I explore the limitless potential of technology to build impactful solutions. My expertise spans full-stack development and artificial intelligence, with a strong foundation in JavaScript, React, Tailwind, Node.js, Express.js, MongoDB, Web3, Ethereum, Ethers.js, Solidity, and Python.
In the AI space, I have worked extensively with machine learning, deep learning, and Generative AI, leveraging frameworks like TensorFlow, PyTorch, Langchain, RAGs, Hugging Face to build intelligent systems. My research and projects involve AI-driven automation, generative models, and data-driven solutions, particularly in healthcare and beyond.
I thrive on tackling complex challenges, continuously expanding my knowledge, and pushing the boundaries of what’s possible. Whether it’s crafting seamless user experiences, optimizing AI models, or exploring the latest advancements in Gen AI, I am always eager to create, learn, and innovate.
Experience
Software Engineer InternRider Levett Bucknall LtdMay 2025 - Aug 2025I developed automation tools to streamline construction document analysis and cost estimation workflows, cutting 5+ hours of manual effort per project. This included automated PDF processing, an intelligent revision tracking system for 1,800+ page specifications, and performance optimization that reduced processing time by 80% (40 minutes to 8 minutes). I deployed these tools on Microsoft Azure with FastAPI, eliminating manual bottlenecks. In parallel, I am building an end-to-end ML pipeline using a Local Classifier per Parent Node (LCPN) to classify scope items into CSI codes, improving F1 accuracy from 84.19% to 92.36% and targeting a 6+ hour reduction per cost estimate. Additionally, I developed a lightweight Python wrapper for parsing PDFs and Excel files, saving 3+ hours in document preparation.Skills:• Python• LLMs• Machine Learning• AuzreOpenAI• Microsoft Azure• Hierarchical Classification• Local Classifier per Parent Node (LCPN)• FastAPI
Graduate Student Researcher [Paper accepted at EMNLP Findings 2025 (acceptace rate: 17%)]University of Southern CaliforniaOct 2024 - May 2025Worked on CulturalPersonas, a research project focused on developing a culturally grounded benchmark for evaluating personality representations in large language models (LLMs). This work addresses the Western bias in traditional personality assessments by leveraging Retrieval-Augmented Generation (RAG) and human validation to create scenario-based, culture-specific evaluations across diverse global contexts. Through this research, I aim to improve AI fairness, cultural adaptability, and personalized AI interactions for a more inclusive and representative AI landscape.Skills:• Python• Generative AI• LLM• RAG• Langchain• OpenAI API• FAISS
Research InternVishwakarma Institute of Information TechnologySep 2023 - Feb 2024Worked on the applications of Deep Learning to auto-design Patient Specific Implants (PSI) using Generative Adversarial Networks (GANs) on MUG500+(B) Dataset. Under the guidance of Dr. Parikshit Mahalle.Skills:• Python• Keras• TensorFlow• NRRD• Computer Vision• Neural Networks• ANTs Image Registration• Data Preprocessing• Jupyter Notebook
Software Developer InternPointsix LabsSep 2022 - May 2023At Pointsix, I developed a React.js-based NFT marketplace, improving user experience and boosting daily active users by 25%. I proposed and implemented an incentivized reward system, attracting 20,000+ users in the first month. I also designed and optimized smart contracts for NFT transactions, reducing gas fees, and built REST APIs that improved response times by 40%. Collaborating with stakeholders, I helped drive user engagement and platform growth.Skills:• ReactJS• Zustand• TailwindCSS• NodeJs• ExpressJS• MongoDB• JavaScript• Solidity• ethers.js• Blockchain• Hardhat
Projects
Want to get in contact with me?
I am currently looking for new opportunities, my inbox is always open. Feel free to reach out to me directly through the below mentioned social media handles for any inquiries, collaborations, or just to say hello! I'm always eager to connect with fellow creatives, clients, and potential collaborators. Whether you have a question or just want to say Hi, I'll try my best to get back to you!