Engineer Redefining Systems with AI-Driven Design

Software Engineer AI Systems & Full-Stack DAU, Gujarat

Mudit Rungta is an innovative Software Engineer & AI Systems Architect passionate about building production-grade AI platforms, intelligent machine learning systems, and resilient full-stack applications. Based in India, he crafts seamless digital experiences accessible globally.

8.02 DAU CPI / 10

600+ DSA Problems

2027 Graduation Year

MUDIT RUNGTA // CORE Active

MUDIT Rungta

Distributed Systems & AI Architecture

Have some questions?

Contact me

NextChapter

Hybrid RAG • PDF.js • 1K+ Users

#1 NextChapter

#2 Graph Fraud GNN

#3 IoT Telematics Platform

#4 50 Cities Cost Index

WORK EXPERIENCE & RESEARCH

FINTECH DISTRIBUTED BACKEND

Software Engineering Intern

BillDesk

Jul 2026 – Aug 2026

Architected a fault-tolerant replication pipeline using Spring Boot & Apache Kafka, streaming 5M+ daily CDC events with partitioned consumers & <150ms latency.
Built a dynamic SQL engine with JSONPath evaluation for multi-table fan-out, automating heterogeneous DB sync to cut manual migration by 85%.
Containerized multi-service test stack (Kafka, MySQL, MongoDB) via Docker Compose and built automated CI/CD pipelines (70% setup time cut).

Spring Boot Apache Kafka (CDC) JSONPath SQL Docker Compose MySQL MongoDB

FINTECH RISK & QUANT RESEARCH

Summer Research Intern

Dhirubhai Ambani University (DAU)

May 2026 – Jun 2026

Researched UPI inclusion patterns for informal workers across India, identifying non-traditional financial behavioral signals.
Engineered an alternative underwriting prototype with a FastAPI risk scoring engine & Web Speech API voice ledger evaluated on 1,200+ merchant cohorts.

FastAPI Python Web Speech API Risk Scoring UPI Analytics

SYSTEMS PORTFOLIO (RANKED BY TECHNICAL IMPACT)

#1 FLAGSHIP AI PLATFORM & HYBRID RAG

Live App ↗ GitHub ↗

NextChapter — AI-Powered Digital Reading & Book Store

Production reading platform with in-book Q&A powered by hybrid vector search (Dense Cosine + Sparse BM25) and Groq Llama 3.1 with metadata-filtered spoiler prevention. Features a virtualized PDF.js reader (<50ms latency) and Pinecone recommendation engine (1,000+ active users, 40% CTR lift).

React 18 FastAPI Groq Llama 3.1 Pinecone BM25 Supabase RLS PDF.js

#2 GRAPH NEURAL NETWORKS & FINANCIAL AI

GitHub ↗

Graph-Based Fraud & Mule Account Detection

Directed topological GNN framework over 6.36M PaySim payment records and 3.28M bank accounts. Engineered structural features (PageRank, clustering coefficient, k-core). Hybrid GAT + XGBoost stacking model reached 75.0% Precision@100 and 0.9129 ROC-AUC with sub-15ms FastAPI inference & PSI drift monitoring.

PyTorch Geometric GAT / GCN XGBoost Stacking cuGraph FastAPI Redis MLflow (PSI)

#3 GCP CLOUD STREAMING & MLOPS

GitHub ↗

Real-Time IoT Telematics Platform

Cloud telematics engine ingesting 10,000+ msg/s sensor streams from 30+ fleet assets via GCP Cloud Pub/Sub. Implemented a dual-tier Feast feature store with BigQuery offline lakehouse (<3ms lookups at zero idle cost) and Optuna LightGBM risk models on Cloud Run via Terraform IaC.

GCP Pub/Sub BigQuery Feast LightGBM Terraform

#4 FASTAPI MICROSERVICE & URBAN ANALYTICS

GitHub ↗

50 Indian Cities Cost Index & Recommender

Urban analytics platform ingesting 50,000+ scraped listings across 50 Indian cities. Delivers sub-50ms priority matching via FastAPI REST endpoints with LRU caching, PCA dimensionality reduction (65% explained variance), and interactive Folium geospatial heatmaps.

Python FastAPI PCA Clustering BeautifulSoup Streamlit

SKILLS & ACADEMICS

Technical Expertise

Languages

Python C++ Java SQL TypeScript / JavaScript C

AI, ML & Graphs

PyTorch Geometric (PyG) GNN / GAT / GraphSAGE LightGBM & XGBoost SentenceTransformers Groq Llama 3.1 Optuna

Data Engineering & Cloud

Spring Boot FastAPI Apache Kafka (CDC) GCP Cloud Run BigQuery Redis Terraform Docker

Education & Achievements

2023 – 2027

B.Tech in Information and Communication Technology

Dhirubhai Ambani University (DAU), formerly DA-IICT

CPI: 8.02 / 10.0

LeetCode 600+ Questions solved; Codeforces Pupil rating.

Qualcomm & IBM Certified: AI Model-to-App Development & ML with Python.

Convener, The Debating Society of DAU (2025–2026).

LICENSES & CERTIFICATIONS

Qualcomm Issued Aug 2026

Show credential

AI Upskilling Certificate: Hands-On Development from Model to App

End-to-end artificial intelligence and machine learning engineering covering deep neural network development, edge model optimization, and full-stack AI application deployment.

Artificial Intelligence (AI) Machine Learning Model Deployment Deep Learning Edge AI

IBM Issued Aug 2026 • ID: 3561837769e245d6af7f3035472a370f

Show credential

Machine Learning with Python

Comprehensive certification in classical ML and statistical learning covering supervised & unsupervised algorithms, regression, classification, clustering, Scikit-Learn pipelines, and model evaluation.

Machine Learning Python Scikit-Learn Supervised Learning Regression & Classification

GET IN TOUCH

Let's build something extraordinary.

I'm currently available for Software Engineering, Backend Data Infrastructure, and AI/ML opportunities.

mudit14127@gmail.com github.com/Mudit-R linkedin.com/in/Mudit-R