Utkarsh Gupta | Developer & Computer Science Student

Tech Stack

(hint: press a key)

My professional journey.

Veermata Jijabai Technological Institute (VJTI)

2024 - 2028

Pursuing a Computer Engineering degree (2024–2028) with a focus on core fundamentals of software engineering.
Actively participating in coding competitions and building full-stack applications.

JavaScript

TypeScript

React

Python

SportIQ (AI-Callable Sports Analytics MCP Server)

May 2026 - Present

Architected and shipped an MCP server exposing 44 AI-callable tools across FIFA World Cup 2026, Formula 1, and IPL cricket, plugging live sports data and predictive analytics directly into Claude, ChatGPT, and IDE assistants with zero setup, keys, or accounts.
Engineered three flagship intelligence tools: a Monte Carlo simulator for the 48-team World Cup bracket on a Poisson xG + Elo engine, an F1 pit-strategy predictor fitting tyre-degradation models to live OpenF1 lap telemetry, and a Dream11 fantasy optimizer solving a real integer linear program via PuLP/CBC.
Designed a resilient FallbackChain on every tool (fresh cache → ordered adapters → stale cache → typed error) with per-source API-budget token buckets and uniform {data, meta} / {error} envelopes, achieving ~95% production tool success at zero infra cost.
Grew to 20,500+ requests and 850+ PyPI installs in ~3 weeks, backed by ~700 hermetic tests (pytest + respx), structured logging, and a fully automated GitHub Actions → Kaniko → GCP Cloud Run deployment pipeline.

Python

Docker

Google Cloud

GitHub

WhisperLearn (AI-Powered Educational Platform)

Jan 2026 - Present

Spearheaded the end-to-end development of an AI study platform, bridging a responsive MERN stack frontend with a complex Python-based machine learning pipeline.
Architected a localized audio-processing engine using OpenAI Whisper and Pydub, automatically segmenting 100MB+ audio files to bypass strict memory limits and eliminate expensive third-party SaaS API costs.
Engineered an intelligent content generator utilizing Groq (Llama 3), KeyBERT, and spaCy to dynamically extract consensus keywords, generating highly accurate contextual summaries, interactive flashcards, and targeted YouTube links.
Designed a seamless, optimistic React UI featuring secure Google OAuth session management, isolated user workspaces, and client-side PDF report generation for study guides.

React

Node.js

Express

MongoDB

Python

InternAgent (AI-Powered Multi-Agent Job Hunter)

Jan 2026 - Present

Architected an end-to-end multi-agent job application platform using the MERN stack and TypeScript, automating job discovery, relevance analysis, and application drafting.
Engineered a robust autonomous pipeline utilizing Playwright and Large Language Models (via OpenRouter/Anthropic), intelligently generating hyper-personalized cover letters and executing dry-run application submissions.
Implemented a resilient job scraping engine with RSS feed fallbacks and hybrid resume-based scoring, ensuring continuous and accurate job matching while circumventing modern anti-bot detections.
Designed a responsive, interactive React user dashboard to monitor agent activities in real-time, manage application queues, and seamlessly handle manual oversight for CAPTCHA challenges.

React

Node.js

MongoDB

TypeScript

Playwright