Himanshu Gupta

Himanshu Gupta

Full Stack AI Engineer

TypeScript / React / Node.js / Python / AI/ML

Open to freelance AI & full-stack projects, as well as full-time roles

Book a CallView Resume

About

8 months of experience. AI & Data Science student, hands-on builder. Shipped real-world AI systems like ClipSync (video pipeline), Medical Scheduling Agent (multi-API), and Math Mentor (multi-agent RAG). Full Stack AI Engineer at Sparkonomy; previously a Python Developer Intern at EspoMedia (OCR, data workflows, model accuracy). I design APIs, integrate models, and ship scalable systems, especially interested in agent workflows and production AI.

Experience

Sparkonomy

May 2026 - Present

Full Stack AI Engineer

Next.jsTypeScriptReactFastAPILangGraphGoogle Cloud

EspoMedia

Jan 2026 - Apr 2026

Python Developer Intern

Tech Department, worked on OCR pipelines, ML/DL model training, and data automation.

  • Assisted in building OCR pipelines using EasyOCR, Tesseract, OpenCV, and preprocessing techniques
  • Supported training and fine-tuning ML/DL models using PyTorch/TensorFlow and YOLO via Roboflow
  • Wrote Python scripts for data cleaning, automation, and dataset preparation
  • Used FFmpeg for video/frame extraction and media processing tasks
  • Worked with MongoDB to store and manage OCR/model outputs
  • Tested model outputs, documented issues, and helped improve accuracy and performance
  • Collaborated with the team, maintained clean code, and followed Git-based workflows
PythonOpenCVEasyOCRTesseractPyTorchTensorFlowYOLOFFmpegMongoDB

Selected Work

ClipSync

AI-powered video editing automation system that intelligently merges A-Roll with contextually relevant B-Roll footage through a single API call. Uses Gemini VLM for visual analysis and implements intelligent placement algorithms for optimal narrative flow.

PythonGemini APIFFmpegNLP

Math Mentor

Freelance · MathPilot
LiveView

Autonomous reasoning system designed to solve high-school and undergraduate level mathematics problems with high reliability. Unlike typical chat interfaces, this application decouples semantic understanding from deterministic computation. The system accepts multimodal inputs (text, image, audio) and employs a Human-in-the-Loop (HITL) workflow to handle ambiguity before it propagates to the solver.

PythonRAGMulti-AgentHITLMultimodal

AI Medical Scheduling Agent

Autonomous conversational AI agent that streamlines patient booking through natural language. Features LangGraph state management, multi-API orchestration with Groq API (Llama 3 70B), and Calendly integration for intelligent appointment scheduling.

PythonLangGraphFastAPIStreamlitGroq API

Stack

Frontend

React.jsNext.jsTypeScriptJavaScriptTailwind CSS

Backend

Node.jsExpress.jsPythonFastAPI

Database

MongoDBPostgreSQLSQLiteSupabase

Languages

PythonJavaScriptTypeScriptC++CSQL

Tools

GitGitHubFFmpegStreamlitPandasNumPy

AI

TensorFlowPyTorchOpenCVLangGraphGroq APIGemini APINLP

GitHub Activity

View Profile
SepOctNovDecJanFebMarAprMayJunJulAug
869 contributions in the last yearLessMore