Utkarsh KumarSoftware Engineer

I build systems that run fast, scale and think smart.

A bit about me

I don't just code with AI, I architect systems where AI does the heavy lifting.

Years in tech
Worked out from your experience dates
Based in
Gurugram, India
Working at
PeopleStrong
Open to
New opportunities in Backend and Generative AI

Where I’ve worked

  1. May 2025 to Now

    Gurugram, Haryana

    SDE-1 (Promoted from Trainee)

    PeopleStrong

    • Optimized large Excel report generation by eliminating memory bottlenecks through JDBC streaming and disk-based writes (Aspose LightCells), using Kafka for request throttling, while accelerating processing with parallel execution.
    • Reduced payslip generation time for 30k employees from 1 hour to 10 minutes by parallelizing PDF generation, optimizing lookups with HashMap, and leveraging GCP Cloud Storage and Cloud Functions.
    • Reduced support-to-developer escalations by building an AI chatbot (Python, Flutter) with a RAG pipeline that answers queries from embedded playbook documentation in Elasticsearch using semantic search.
    • Implemented Multi-Factor Authentication (MFA) for a legacy ASP.NET Web Forms app by integrating a Spring Boot OTP service for email and SMS delivery, supporting OTP expiry, resend limits, and account lockouts using SQL Server.
    • Developed a Payslip Builder using Spring Boot, SQL, and Flutter, enabling configurable payslip layouts with customizable field ordering and element alignment.
    • Mentored trainees in building an AI-powered payroll assistant using Python, LangGraph, MCP, RAG and React to answer employee payroll queries through natural language.
    • Built an automated reporting API to email Jenkins SQL pipeline execution logs from the past 24 hours and scheduled it using Rundeck cron jobs.
    • Developed Spring Boot APIs for uploading and managing company logos in Google Cloud Storage (GCS) using the Google Cloud SDK.
  2. Jan 2025 to May 2025

    Remote

    Software Engineer Intern

    Ionage

    • Implemented Razorpay refund workflow with webhook-based status tracking in the company’s wallet system using Golang, enabling reliable and automated payment refunds.
    • Enabled fast vehicle lookups by implementing a read-through cache.
  3. Jan 2024 to Jul 2024

    Remote

    Software Engineer Intern

    Clienter AI

    • Digitized timesheet submission and payroll calculation by building an end-to-end module with Spring Boot and Next.js (TypeScript) for accountants.
    • Improved app responsiveness by implementing paginated APIs (Spring Data JPA) and frontend debouncing to reduce unnecessary server requests.
    • Utilized Static Site Generation (SSG) to create SEO-friendly blog pages using Markdown files.

Products I’ve built

Integrated Architecture Environment

  • A real-time collaborative system-design platform built on tldraw and Yjs, with live sync and role-based project sharing.
  • AI-assisted architecture generation and markdown spec export using Groq and Trigger.dev.
  • WebMCP lets AI agents use the app through browser-native tools.
  • Next.js
  • Express.js
  • PostgreSQL
  • Yjs
  • tldraw
  • WebMCP
  • Trigger.dev
Log2PR architecture diagram

Log2PR

Elasticsearch Hackathon

  • A multi-agent AI system that pulls error logs from Elasticsearch and turns them into code fixes.
  • Auto-generates GitHub PRs for each fix.
  • Notifies developers via Slack.
  • LangGraph
  • LangChain
  • Model Context Protocol
  • Deep Agents
SemantiPay screenshot

SemantiPay

TiDB Hackathon

  • A payroll system that adapts instantly to compliance changes using semantic search and agentic workflows.
  • Payroll calculation with LLMs, using Jina AI for embeddings and Groq for inference.
  • Dify AI workflows alert HR by email and Slack on compliance violations.
  • TiDB
  • Dify AI
  • Express.js
  • Next.js
  • Upstash Redis
  • Jina AI
  • Groq
TrendyThumbs screenshot

TrendyThumbs

  • Analyzes thumbnails of trending YouTube videos with Google Vision AI: dominant colors, word count, expressions and object labels.
  • Delivers insights through shadcn/ui charts so creators can spot design trends.
  • Spring Boot
  • Next.js
  • MongoDB
  • shadcn/ui
  • Magic UI
  • Aceternity UI
  • Caffeine Caching
  • Docker

What I work with

  • Languages

    Java, Python, JavaScript, TypeScript, SQL

  • Backend and infra

    Spring Boot, Express.js, FastAPI, Git, Docker, MongoDB, Redis, Kafka, AWS

  • Gen AI

    LangGraph, LangChain, RAG, Model Context Protocol, Agentic AI, Vector Search

  • Frontend

    React, Next.js, Flutter, React Native

Certificates I hold

CertificationIssued byYear
Generative and Agentic AI with PythonUdemy2025
Java Spring Framework 6 with Spring Boot 3Udemy2024
Flutter MasterclassNet Ninja2025

Leadership and extracurriculars

  1. Nov 2023 to May 2025

    Surathkal, Karnataka

    Executive Member

    Google Developer Groups, NITK Surathkal

    • Led the development of the official website for NITK Photography Club.
    • Developed the UI for an NITK exclusive Open Source Archive on the COSH website, attracting over 10k visits.

Where I studied

  1. Dec 2021 to May 2025

    Surathkal, Karnataka

    Bachelor of Technology

    National Institute of Technology

  2. 2021

    Indore, Madhya Pradesh

    Class XII

    The New Green Field Public Academy

  3. 2019

    Bina, Madhya Pradesh

    Class X

    DAV BORL Public School