🚀 Full-Stack Developer & AI Systems Builder | Top 7 @ HackFest 2 among 500+ | AI/ML Intern @ Infosys | Shipping Production-Grade Apps
-
🔭 I'm currently working on CareerTrack — a full-stack job application tracker with AI-powered resume-to-JD gap analysis using Groq LLM
-
🏆 Achieved Top 7 at HackFest 2, GDG Cloud New Delhi among 500+ participants with DataSense AI — a multi-agent data dictionary generator
-
🌱 I'm currently learning Advanced System Design · LLM Orchestration & Multi-Agent Pipelines · PostgreSQL & Supabase at scale
-
💼 Past experience: AI/ML Intern @ Infosys Springboard — built production-grade NIDS using ensemble ML models (Decision Trees, Random Forests, Gradient Boosting)
-
👯 I'm looking to collaborate on Notes-App
-
📝 I regularly write articles on medium.com/@upasanaprabhakar35
-
💬 Ask me about MERN Stack, Full-Stack Architecture, Multi-Agent AI Systems, REST APIs, ML Engineering
-
📫 How to reach me: upasanaprabhakar35@gmail.com
React · Node.js · PostgreSQL · Prisma · Socket.io · Groq LLM · JWT Auth
- Built and deployed a production-grade job tracking system with 7-stage pipelines, real-time notifications via Socket.io, and AI-powered resume-to-JD gap analysis using Groq LLM
- Reduced manual interview prep time by 60% through role-specific AI suggestions and keyword gap detection
- Implemented JWT auth with httpOnly refresh token rotation, rate limiting, and CORS across Vercel + Render + Neon PostgreSQL
React · Node.js · Groq (Llama 3.3 70B) · Supabase · D3.js
- Built a 3-agent pipeline (Atlas → Sage → Guardian) that turns any undocumented CSV or database into a fully documented data dictionary in under 30 seconds
- Guardian auto-scores data quality (0–100) and detects PII across 11 categories with GDPR risk classification
- 🏆 Top 7 at HackFest 2, GDG Cloud New Delhi among 500+ participants
Python · Scikit-learn · SMOTE · Pandas · NumPy
- Built production-grade NIDS using ensemble ML models achieving high-accuracy threat detection
- Reduced false positives by 25% and improved detection speed by 30% via SMOTE-based feature engineering
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
- 🥇 Top 7 — HackFest 2, GDG Cloud New Delhi | Multi-Agent AI pipeline among 500+ participants (Feb 2026)
- 📜 Stanford University Online — Divide and Conquer, Sorting and Searching, and Randomized Algorithms (Feb 2025)
- ☁️ Google Cloud Skill Badge — Develop GenAI Apps with Gemini and Streamlit (Aug 2025)