Full-Stack & AI Engineer

dallas, tx · cdt

Full-StackAI / LLMMulti-AgentReact · Next.jsPython

I build full-stack products

B.S. CS @ UT Dallas, Dec 2026 · open to internships & full-time roles

AI Society UTD · Tech Officer|Syft · SWE Intern|Nebula Labs · Project Lead|Glenart Group · Capstone|Maiya LLC · SWE Intern

01

About

Available Dec 2026

I'm Harsh Patel, a Computer Science student at UT Dallas and a full-stack engineer working at the intersection of web development and AI. I build React and Next.js front ends, Python and Node back ends, and increasingly, multi-agent LLM systems with Claude and OpenAI. I've shipped software for platforms serving 500+ and 20,000+ users, led a 30-person engineering team, and won first place at TAMUHack's Toyota Challenge.

I treat AI agents as real engineering leverage.

— with a heavy emphasis on prompt engineering and validating what the model actually produces.

B.S. Computer Science

UT Dallas · GPA 3.63 · Aug 2023 – Dec 2026

LanguagesTypeScript · JavaScript · Python · Java · SQL · C++ · HTML · CSS
AI / LLMLLM APIs (Claude · OpenAI · Hugging Face) · Prompt Engineering · Prompt Chaining · Agentic Workflows · Multi-Agent Orchestration · LangChain · CrewAI · RAG · Semantic Search · Model Fine-Tuning · Generative AI
Frontend & ToolingReact · Next.js · Vue.js · Tailwind CSS · Framer Motion · Recharts · Figma · Front-End Integration · a11y
Backend & DataNode.js · Express · Flask · FastAPI · REST APIs · WebSockets/SSE · Asynchronous Programming · Vector Databases (pgvector) · PostgreSQL · MySQL · MongoDB · Supabase · Firebase
Cloud & DevOpsAWS (S3 · Lambda) · GCP · Railway · Docker · Git/GitHub · CI/CD (GitHub Actions) · Linux

Featured Work · Industry Capstone

DOME — Glenart Group

A 13-agent Claude API pipeline that extracts structured equipment data from single-line electrical diagrams — cutting a 2–4 hour manual task to under 5 minutes.

95%

Equipment-level accuracy

5 min

Down from 2–4 hours

13

Agents in the pipeline

Source code confidential · Sponsor: Glenart Group · Spring 2026

Built with

  • TypeScript
  • React
  • Next.js
  • Python
  • Node.js
  • FastAPI
  • Flask
  • PostgreSQL
  • Supabase
02

Projects

Five builds, four awards · click any row for details

01/05

RideIQ

→

🥇 1st Place — Toyota Challenge @ TAMUHack

AI · Full-Stack · Hackathonteam · hackathon

Led front-end for an AI-powered car marketplace built in 24 hours. pgvector semantic search over 43k vehicles and 62k listings. 4-car comparison with Recharts. Test-drive booking flow with Resend.

1st place. Full working product in 24 hours.

ReactNext.jsTypeScriptpgvectorSupabaseRechartsTailwind

03

Experience

  1. 30%

    front-end efficiency boost

    Technology Officer

    Current

    Artificial Intelligence Society · Jun 2025 – Present

    Built and maintained the org's web platform serving active members across UT Dallas. Engineered a reusable React component library that cut feature development time by 30%. Defined front-end standards for consistency across concurrent projects and enabled non-technical officers to manage content independently through documentation and internal tooling.

    Stack —React · Next.js · TypeScript · Tailwind

  2. 20,000

    users on the redesigned platform

    Project Lead

    Leadership

    Nebula Labs — UTD Trends · Jan 2025 – May 2025

    Owned end-to-end delivery of a complete platform redesign for UTDTrends, a course-aggregation tool used by 20,000 UT Dallas students. Coordinated 30+ engineers and designers across parallel workstreams — setting sprint cycles, reviewing PRs, and unblocking the team. Brought the platform from design to production in 5 weeks and established CI/CD pipelines with GitHub Actions to automate deployment and reduce manual release overhead.

    Stack —React · GitHub Actions · CI/CD

  3. 500+

    active users supported

    Software Engineer Intern

    Internship

    Syft (Startup) · Sept 2024 – July 2025

    Joined an early-stage fashion-tech startup building AI-powered trend categorization tools. Developed and shipped features for an internal Flask web app used by 500+ active users to classify product images into defined fashion trend categories. Automated the bulk image ingestion pipeline — replacing a manual upload process with an AWS S3 + Firestore integration via Boto3, processing 1,000+ images per run.

    Stack —Python · Flask · AWS S3 · Firestore · Boto3 GitHub →

  4. 30%

    reduction in check-in wait times

    Software Engineer Intern

    Internship

    Maiya LLC · Sept 2023 – March 2024

    Built a full-stack property management system for a motel business serving 500+ monthly guests — replacing manual processes with a unified dashboard for bookings, reservations, and financial tracking. Designed the check-in/check-out flow in React with a Node.js/PostgreSQL backend, cutting average wait times by 30%. Built scalable REST endpoints that automated data sync across front desk, billing, and reporting systems.

    Stack —React · Node.js · Express · PostgreSQL

by the numbers.

0+

Award-winning projects

0+

Users reached

0

Internships

0st

Place hackathon finish