Incoming SWE Intern @ Microsoft

Computer Science Major @ University of Utah
Incoming SWE Intern @ Microsoft | AI Intern @ UIT | CS @ University of Utah
🌐 Portfolio • LinkedIn • Wingman.ai • 📍 Salt Lake City, UT

I’m an early-career AI developer building full-stack applications at the intersection of AI, cloud, and healthcare/legal tech.
From HIPAA-compliant hospital LLM interfaces to legal AI benchmarks, I specialize in solving real-world problems with high-performance AI systems.

Jan 2025 – Present

• Co-built a secure, HIPAA-compliant chat platform for hospital staff to interact with LLMs grounded on internal data
• Built React-based frontend and scalable AWS backends (Lambda, S3, Bedrock) for dynamic LLM selection and file uploads
• Reduced LLM inference latency by 40% and improved retrieval query speed by 60% via caching + DB optimizations
• Enabled real-time interactions with hospital data through a secure, fault-tolerant infrastructure

Aug 2025 – Present Collaborations with NASA, MIT, Microsoft, DoD

• Advancing AI-driven materials discovery by developing ML pipelines for high-throughput experimentation, computational modeling, and aerospace materials design.
• Integrating large language models (LLMs) and multi-agent AI systems to streamline knowledge sharing, automate literature mining, and enhance decision support across interdisciplinary research teams.
• Collaborating with cross-domain experts in computer science and materials engineering to design AI workflows that accelerate research in rocket engines, hypersonics, and in-space manufacturing.

Nov 2024 – Apr 2025

• Processed and structured 10M+ Indian legal documents across statutes, orders, and metadata
• Built task-specific RAG pipelines improving retrieval accuracy by 28% and reducing hallucinations by 35%
• Benchmarked LLMs (InLegalBERT, InLegalLLaMA, GPT-4o-mini) across 20+ tasks using LegalBench, NyayaAnumana, etc.
• Synthesized 15+ research papers into internal strategy docs for legal AI design and evaluation

React Native, SwiftUI, OpenAI API, Firebase

• Developed multi-modal AI assistant using GPT-4o and Whisper to process chat, voice, and image input
• Integrated Google Calendar, secure Firebase Auth, and real-time data sync
• Built MVVM-based SwiftUI interface with Combine for reactive state management
• Supports 1000+ daily users with offline capabilities and personalized memory recall (RAG)

Python, NebulaGraph, LlamaIndex, Docker, Chainlit

• Designed and deployed a distributed GraphRAG platform querying 1M+ biomedical entities
• Integrated UniProt, AlphaFold, RXNav into ETL pipeline with deduplication + validation
• Reduced query latency 3x via high-degree node analysis, caching, and graph schema refinement
• Powered explainable, up-to-date responses by augmenting LLMs with structured graph data

SwiftUI, Flask, PostgreSQL, Docker, Ollama (DeepSeek 7B)

• Built AI-driven iOS app for querying 50+ rare diseases and visualizing drug repurposing recommendations
• Fine-tuned DeepSeek 7B on biomedical data, achieving 92% accuracy in drug-disease matching
• Integrated data from DrugBank, PubMed, ClinicalTrials.gov, etc. with robust REST APIs
• Designed for real-time usage by researchers and clinicians, emphasizing security and speed

Languages: Python, Java, C++, C#, JavaScript, TypeScript, Swift, SQL
Frameworks: React, Flask, Node.js
AI/ML: OpenAI API, Ollama, LlamaIndex, LangChain, TensorFlow, RAG/GraphRAG, Multi-Agent Systems, MLflow
Cloud: AWS (Lambda, S3, RDS, DynamoDB, Bedrock), Firebase, Docker
Databases: PostgreSQL, Vector Stores, NebulaGraph

University of Utah, Salt Lake City, UT
B.S. in Computer Science, Aug 2023 – May 2027
GPA: 3.7 / 4.0
Relevant Coursework: Machine Learning, NLP, Computer Vision, Data Structures & Algorithms, DB Systems

• 🧠 Cloud AI Systems
• ⚙️ Retrieval-Augmented Generation (RAG)
• 🤖 Multi-Agent AI
• 🔬 Bioinformatics & Legal AI
• 🎯 Research-driven development

• ✉️ Email: edu.rathodkunj@gmail.com
• 🔗 LinkedIn: linkedin.com/in/rathodkunj
• 💻 GitHub: github.com/rathodkunj2005