Pratham Gupta

I build production AI agents and ML systems that ship.

ML Intern at Tower Research Capital · previously AI at Inteligenai · ICPC Asia West Regionalist · GATE AIR 104

From multi-agent systems and RAG pipelines to low-latency model serving and full-stack web apps, I take AI products from idea to production.

Experience

  1. Tower Research Capital

    ML Intern · Core AI & Machine Learning

    Gurgaon, India

    Jan 2026 – Present
    • Build auto-research tooling that speeds up quantitative research workflows.
    • Develop proactive and reactive AI agents, contributing to the zero-to-one build-out of internal agentic workflows that automate firm-wide operational tasks.
    AI AgentsQuant Research ToolingAutomation
  2. Inteligenai

    AI Intern

    Gurgaon, India

    Jul 2024 – Dec 2025
    • Designed the end-to-end architecture and delivered multiple large-scale AI solutions across diverse client industries.
    • Engineered the AI backend for Spector.ai, an AI-first industrial agent platform; architected a multi-agent system using LangGraph for automated root-cause analysis.
    • Deployed Vision Transformer (ViT) models on Azure Functions with ONNX optimization and quantization, cutting latency from 1s to 100ms and scaling to 1B+ inferences a year.
    • Built an enterprise-scale RAG pipeline that parses complex formats (PDF, DOCX) with HNSW vector search, integrating LLMs via LangChain for semantic reasoning.
    • Set up CI/CD with GitHub Actions and Azure DevOps to automate deployments on Docker and Kubernetes.
    LangGraphRAGViTONNXAzureKubernetes

What I build

The kinds of systems I work on, from AI agents to full products.

AI Agents & Automation

Multi-agent systems that research, reason and act across tools, turning repetitive operational work into automated workflows.

LangGraph multi-agent system for automated root-cause analysis at Spector.ai

RAG & Knowledge Assistants

Assistants that answer questions from documents and internal knowledge, with answers grounded in the source material.

Enterprise-scale RAG pipeline over PDF and DOCX documents

Model Deployment & Optimization

Taking models from notebook to production, serverless or containerised, tuned for latency and cost.

Vision Transformer models on Azure Functions, optimised with ONNX and quantization

Full-Stack Web Apps

Fast, reliable web products with React / Next.js frontends and Python backends that hold up under traffic spikes.

CBSE results platform: 125K+ requests at peak, 5,000+ schools

Working on something along these lines? I'd love to hear about it

Featured Projects

CBSE Results Parser & Analytics Platform

Next.jsFirebaseCloudflarePython

A web platform that turns raw CBSE result files into school-level analytics. It ran under heavy load during result hours, with sub-second responses throughout.

5,000+
Schools
125K+
Peak requests
10K+
Unique users
  • Custom parsing algorithm with built-in statistical analysis
  • Cloudflare caching and DDoS protection to absorb peak-hour traffic
  • Web and CLI interfaces, with Firebase authentication
Live site

TripAlly: AI Travel Assistant

PythonLangGraphFastAPIGroq

A multi-agent AI travel concierge that browses the web on its own to plan trips and carry out bookings end to end.

  • Multi-agent architecture running on Groq's low-latency LLMs
  • Autonomous web browsing for search and booking
  • Server-Sent Events stream each agent step to the user in real time
Source on GitHub

Awards & Competitions

ACM ICPC

Asia West Regionalist

Qualified for regionals twice, in 2024 and 2025

GATE DS & AI

All India Rank 104

Among 91,000+ candidates

HackRX 6.0 · Bajaj Finserv

Winner, Darkhorse Award

Top 4 of 10,000+ participants

Technical Skills

Languages
PythonC/C++JavaScript
AI / ML
LangChainLangGraphRAGMulti-Agent SystemsComputer VisionONNXModel Fine-tuningNumPypandas
Web & Backend
FastAPIDjangoFlaskReact.jsNext.jsAstro
Cloud & DevOps
Microsoft AzureGoogle CloudDockerKubernetesGitHub ActionsFirebaseCloudflareLinuxGit

Education

Indian Institute of Technology, Madras

Bachelor of Science in Data Science and Applications

2023 – 2027 (Expected)

CGPA

9.53

Maharaja Agrasen Institute of Technology

Bachelor of Technology in Computer Science Technology

2023 – 2027 (Expected)

CGPA

9.45

Get in touch

Have an interesting idea or problem in AI, or just want to talk shop? My inbox is always open.

[email protected]