Shresth Bhargava
Available for new opportunities

Building systems that scale intelligently.

I am a backend engineer focused on resilient architecture, AI-driven applications, and robust APIs. Engineering real software to solve complex problems.

Scroll

Currently focusing on

Featured Work

Engineering Case Studies

AI SystemsFeatured

CivicOS

An AI-powered civic intelligence platform that maps citizen issues to responsible government departments, officials, laws, and complaint channels.

Java 21Spring BootPostgreSQLRedisKafkapgvector
API latency reduced 97%
Recursive CTEs
Spring Boot
Redis
Kafka
Backend

Vellum (AutoResearch AI)

Rebuilt the backend of an AI startup documentation platform from Python to Java Spring Boot. Layered architecture with a RAG-based multi-agent pipeline.

Java 21Spring BootPostgreSQLJWTRAG
RAG Architecture
Multi-agent pipeline
Java 21
Spring Boot
PostgreSQL
Full Stack

HandymanHub

Full-stack service marketplace connecting customers with workers through skill and location-based discovery.

Java 17Spring BootMySQLNext.jsDocker
N+1 queries fixed (19 to 1)
29 Unit Tests
Next.js
Spring Boot
Docker
AI Systems

Loan Eligibility Predictor

Machine learning classification system to predict loan approval eligibility based on applicant financial attributes.

PythonScikit-LearnStreamlitPandas
SHAP-based explainability
Random Forest
Python
Scikit-Learn
Streamlit

Deep Dive

CivicOS Architecture

A look into the jurisdiction-aware AI mapping platform. Engineered for scale, speed, and strict regional correctness without LLM hallucinations.

01

API Gateway

Routes requests, handles JWT stateless authentication, and enforces rate limiting to protect the downstream services.

02

Spring Boot Core

Executes business logic using constructor-based dependency injection. Interfaces directly with Kafka for event streaming and search analytics.

03

Redis Caching

Caches frequent regional queries, reducing latency from 900ms to 27ms. Uses TTL and cache invalidation strategies for fresh data.

04

pgvector + Ollama RAG

PostgreSQL handles recursive CTEs for accountability mapping. pgvector enables semantic search fallback utilizing an integrated LLM pipeline.

Technical Skills

The Stack

Technologies I use to build scalable, secure, and highly performant backend systems and AI architectures.

Backend Engineering

Spring BootSpring SecurityREST APIsHibernateJPAJWTFastAPINode.jsExpress.js

AI Engineering

LLM APIsRAGVector SearchpgvectorLangChainLangGraphGemini APIOllama

Databases

PostgreSQLMySQLMongoDBRedisFlyway

Cloud & DevOps

DockerDocker ComposeGitHub ActionsCI/CDRenderVercelSupabaseLinux

Languages & Core

JavaPythonJavaScriptSQLC++

Testing & Tools

JUnit 5MockitoTestcontainersGitGitHubMavenPostmanSwagger/OpenAPIIntelliJ IDEA

Experience

Open Source & Education

Open Source Contributor

GirlScript Summer of Code (GSSoC) 2026
May 2026 - Present

Merged 47+ pull requests across open-source repositories with 4,445+ contribution points. Delivered feature enhancements, bug fixes, and accessibility improvements. Collaborated with maintainers through GitHub issues, reviews, and workflows.

GitOpen SourceIssue-driven Development

Open Source Contributor

Elite Coders Summer of Code (ESoC)
June 2026 - Present

Selected as an ESoC 2026 contributor for production-oriented open-source development. Focused on collaborative software engineering practices and feature implementation.

Git WorkflowsCode Reviews

Open Source Contributor

Nexus Spring of Code (NSoC'26)
June 2026

Ranked among top contributors (228) with 7 merged pull requests and 53+ contribution points. Contributed improvements across accessibility, security, UI/UX, and performance optimization.

UI/UXSecurityPerformance Optimization

B.Tech Computer Science Engineering

VIT Bhopal University
Aug 2024 - Aug 2028

Currently pursuing a Bachelor of Technology. Focusing heavily on backend engineering, data structures, and building real-world projects outside the classroom.

Data StructuresAlgorithmsJava

About Me

Engineering with correctness.

I fixed an API that took 900ms to respond. It now takes 27ms.

That's not a resume line — that's what happens when you stop treating backend work as "make it function" and start treating it as "make it correct under load."

I am a backend-focused CS student at VIT Bhopal ('28). Most of what I know didn't come from lectures. It came from using Redis caching, recursive PostgreSQL CTEs, and designing jurisdiction-aware search on CivicOS. It came from cutting database calls from 19 to 1 to fix an N+1 query problem on HandymanHub.

Currently, I am working with Java, Spring Boot, PostgreSQL, Redis, Docker, and increasingly in RAG/LLM-based systems (LangChain, pgvector, Gemini API). I like backend problems where correctness and performance both matter.

As an open-source contributor, I have merged over 47+ pull requests into production repositories (GSSoC 2026), ranking among the top contributors program-wide.

47+
OSS PRs Merged
97%
Latency Reduction
1.34L+
Amazon ML Shortlist
290+
DSA Problems