Hi, I'm Purav Parekh

    Full-Stack AI Engineer

    I build AI agents, multi-agent orchestration systems, and full-stack web applications. Specializing in LLMs, RAG pipelines, and scalable React/Node.js backends — from health-tech startups to NIST research.

    Purav Parekh LinkedIn profilePurav Parekh GitHub profilePurav Parekh Medium blog
    Purav Parekh - Full-Stack AI Engineer

    Work Experience

    A journey through my professional career building innovative solutions

    Founding AI Engineer

    Stealth Startup (Health-tech)

    Aug 2025 - Present
    Remote, USA
    • Architected a HIPAA-compliant PII redaction layer using Google DLP, detecting and redacting sensitive data from all user-uploaded documents and real-time conversations before storage, achieving 98% automated redaction with zero manual intervention.
    • Designed a multi-agent orchestration system for intelligent request routing replacing hardcoded model calls with a pipeline of specialized agents handling query classification, selective data fetching, medical reasoning, response summarization, and proactive appointment preparation.
    • Built a long-term memory layer using Mem0, enabling the AI to retain and recall patient context across sessions, eliminating repetitive user inputs and making interactions feel continuous and personalized
    • Reduced LLM token processing costs by ~20% through custom filtering logic in the Data Fetching Agent, indexing only high-value medical context and discarding conversational noise.
    • Engineered a bidirectional sync pipeline that instantly purges memory storage when users edit/delete health records, ensuring zero data staleness and strictly adhering to data sovereignty requirements.
    Agents
    Agent Orchestration
    OpenAI
    Mem0
    React
    TypeScript
    Firestore

    Senior Full-Stack Engineer

    National Institute of Standards and Technology (NIST)

    Oct 2023 - Present
    Maryland, USA
    • Led the design and architecture of the first-ever pipeline to digitize unstructured sustainability compliance documents, leveraging context engineering, prompt inject defenses, and LLM-driven schema inference to extract and validate structured product data from raw PDF/HTML documents.
    • Built automated AI-powered data pipelines with Azure Document Intelligence and Blob Storage SDKs, converting 1,000+ PDFs into structured CSVs and reducing processing time by 95%, saving ~$1,000/month in labor.
    • Modernized a 40-year old building economics analyses tool used by the US military, migrating a Java desktop executable to a performant React web application with RxJS state management
    • Designed and optimized PostgreSQL database schemas and indexing strategies, improving query response time by 25%, boosting app performance by 20%.
    • Developed and deployed RESTful APIs in Node.js, enhancing data throughput and reducing API latency by 15% through efficient caching and query optimization.
    LLM
    LangChain
    OpenAI
    Pinecone
    React
    TypeScript
    Node.js
    PostgreSQL

    Software Engineering Intern

    Normalyze Inc

    May 2022 - Aug 2022
    California, USA
    • Refactored a tightly coupled, inconsistently structured frontend into a modular React component architecture reducing technical debt by 40% and improving development speed for the engineering team.
    • Achieved 95% unit and functional test coverage using Mocha.js and Chai.js, ensuring high code reliability across mission-critical features.
    • Built custom admin tools using Retool, automating manual workflows for the Customer Success team and increasing productivity by 70%.
    • Deployed the tool on AWS EC2 and integrated with REST APIs and Amazon RDS (PostgreSQL) for secure, scalable data access and management.
    React
    Express.js
    Node.js
    PostgreSQL
    Docker
    AWS

    Software Engineer

    Adcuratio Media

    Jan 2021 - Aug 2021
    Remote, India
    • Accelerated development by creating a reusable React component library integrated with MobX state management, reducing manual contract work by 50%
    • Integrated RESTful APIs with React using hooks and MobX for state management, enhancing application responsiveness and user experience.
    • Optimized data handling and API efficiency by implementing virtualized rendering and reducing HTTP requests by 60% through API call batching.
    Javascript
    React
    CSS
    PostgreSQL
    Docker

    Founder

    Parekh Digital Solutions

    March 2019 - Aug 2021
    Remote, India
    • Founded an information technology company delivering end-to-end solutions—web development, e-commerce, SEO, and business strategy—for clients across India, including founders, CEOs, and small business owners.
    • Designed and deployed full-stack web applications using React and Django, driving up to 15% increases in client revenue and user engagement.
    • Implemented robust debugging workflows, database optimizations, and frontend performance improvements—reducing issue resolution time by 25% and enhancing system reliability.
    • Made critical technical and product decisions that directly impacted client profitability, scalability, and long-term success.
    JavaScript
    React
    Python
    Django
    PostgreSQL

    Skills & Expertise

    Technologies and methodologies I use to build amazing products

    Web Development

    React
    Javascript
    TypeScript
    HTML/CSS
    Tailwind CSS

    AI & Machine Learning

    AI Agents
    Multi-Agent Orchestration
    Large Language Models (LLMs)
    Generative AI
    RAG
    OpenAI
    LangChain
    Pinecone
    Mem0
    Python

    Frameworks & Libraries

    Node.js
    Express
    FastAPI
    Django
    Electron
    Vite

    Programming Languages

    JavaScript
    TypeScript
    Python
    Java

    Cloud & Databases

    AWS
    Azure
    PostgreSQL
    MongoDB
    Redis
    Docker
    Kubernetes

    Tools & Methodologies

    Git
    CI/CD
    Agile
    Testing
    Monitoring
    Security
    Performance

    Featured Projects

    A selection of projects that showcase my technical skills and creativity

    View all on Github →

    Cricket League RAG Chat

    A RAG-powered application that uses LLMs+Knowledge base to answer questions about the laws of the local cricket league.

    RAG
    OpenAI
    Pinecone
    React

    parsEPD

    A LLM based tool that helps digitize Environmental Product Declarations (EPDs).

    LLMs
    OpenAI
    React
    TypeScript

    Browser Notes

    A chrome extension that enables users to save notes in the browser.

    Javascript

    Movie App

    A Nuxt.js web app that shows real-time trending movies & tv shows across genres with ratings.

    Nuxt.js
    Vue.js
    JavaScript

    Code2Learn

    A user-friendly web application in Django that provides non-technical users with access to community-curated project-based tutorials for learning programming languages & tools.

    JavaScript
    Python
    Django

    Medovation Inc

    Full stack Django application displaying company’s products & services, that could be searched, ordered, filtered & sorted based on user preferences.

    JavaScript
    Python
    Django
    SQL

    Certifications

    Professional certifications that validate my expertise

    oracle

    OCI Gen AI Professional

    Oracle
    2025
    Oracle
    Generative AI
    LLM
    amazon web services

    AWS Certified AI Practitioner

    Amazon Web Services
    2024
    AWS
    AI
    Cloud Architecture

    Publications

    Research and insights shared with the developer community

    View all on Google Scholar →

    Environmental Product Declaration (EPD) Database Creation and Public Life Cycle Inventory (LCI) Data Gap Assessment

    2025

    Examining how artificial intelligence can enhance user experience through personalization, predictive interfaces, and intelligent automation...

    Python
    Azure
    Automation
    Sustainaibilty

    Energy Price Indices & Discount Factors for Life-Cycle Cost Analysis

    2025

    An in-depth analysis of React performance optimization techniques, covering everything from bundle splitting to server-side rendering...

    Python
    Automation
    Economics

    Building Life Cycle Cost (BLCC) User Guide

    2025

    This paper explores the integration of machine learning models into web applications while maintaining performance and scalability...

    React
    Web Development
    Economics

    A Guide to the Smart Investment Tools

    2024

    This paper explores the integration of machine learning models into web applications while maintaining performance and scalability...

    React
    Web Development
    Economics

    Content & Media

    Articles, podcasts, and videos where I share knowledge and insights

    Medium • 2025

    Tokens, Context, and Embeddings: The Secret Language of LLMs

    A comprehensive guide to the fundamental jargon of Large Language Models.

    LLMs
    Tokens
    Embeddings
    YouTube • 2024

    GWU to NIST: Purav Parekh's American Quest

    A podcast where I share my journey in the United States as a graduate student and landing a job at NIST, with the ISTP Magazine

    Student
    Interviews
    NIST
    YouTube • 2022

    Dynamic Webpage Search with AJAX

    Presentation at GW Coders about a quick 30 minute deep dive into Webpage search using AJAX.

    Javascript
    AJAX
    Tutorial
    Medium • 2025

    From Regex to AI: Engineering a scalable Document Parsing Pipeline

    A detailed blog on how we automated our document parsing pipeline and the lessons we learnt.

    Azure
    Machine Learning
    Parsing
    Automation
    Medium • 2025

    Attention is All You Need — Here's What That Really Means

    Explaining the concept - Attention Transformer, that drives modern day LLMs in the simplest way possible

    Transformer
    LLMs
    AI
    Medium • 2023

    Why Message Queues?

    A deep dive into what are message queues and why we need them.

    System Design
    Cloud
    Distirbuted Systems

    Get In Touch

    Let's discuss opportunities, collaborations, or just say hello

    Direct Contact

    Prefer direct communication? Feel free to reach out via email for immediate response.

    Response Time

    I typically respond to messages within 24-48 hours. For urgent matters, please use the direct email option above.

    Send a Message