Gaurav Mehta.

Full-Stack Software Engineer

Open to software engineering roles · Graduating June 2026

Focused on end-to-end product development with TypeScript, Next.js, Node.js, and PostgreSQL. I prioritize performance, security, and maintainable code.

ResumeGet in touch

Experience

Full-Stack Intern · Second Brain Labs

Nov 2025 - Dec 2025
  • Built an end-to-end AI-driven campaign suggestion pipeline using LinkedIn data extraction and Apify APIs to generate personalized outreach campaigns.
  • Integrated an LLM-based content generation pipeline for dynamic message personalization, reducing manual effort and improving engagement quality.
  • Enhanced manually created campaigns with LLM-assisted workflows — refining messaging, clarity, and personalization.
  • Developed external APIs for campaign management, automated outreach, and user lifecycle handling.

Software Engineering Intern · Trading Studio

Apr 2025 - Aug 2025
  • Optimized PostgreSQL queries and indexing, reducing response times from 3s to under 500ms for core analytics workflows.
  • Increased test coverage from 40% to 85% with an automated testing framework, significantly reducing production bugs.
  • Delivered 60+ production features and improvements within 4 months.
  • Improved system responsiveness by up to 600% in high-traffic areas.
  • Built a query management platform enabling non-technical users to create and analyze complex reports.

Featured Work

NextFlow screenshot 1
NextFlow screenshot 2
NextFlow screenshot 3

NextFlow

Shipped

Visual LLM workflow builder — a drag-and-connect canvas where workflows execute as recursive DAGs with live per-node state streamed to the browser.

  • Architected a recursive DAG execution engine on Trigger.dev where each node blocks only on its direct upstream dependencies — fast branches proceed immediately instead of stalling behind slow siblings.
  • Guaranteed exactly-once execution in diamond-shaped DAGs with scoped idempotency keys, so shared upstream nodes run a single time even when fanned out to concurrent downstreams.
  • Eliminated client polling by streaming live per-node execution state over a single SSE channel per workflow, cutting redundant serverless invocations.
  • Used checkpointed waits so time-delayed nodes consume zero compute while suspended; the database is history-of-record, not a coordination layer.
  • Built end-to-end: React Flow canvas with debounced autosave, Clerk-authenticated + Zod-validated API routes, and Gemini/Transloadit integrations.
Next.js 14TypeScriptTrigger.dev v4React FlowPostgreSQLPrismaClerk
InstantPay screenshot 1
InstantPay screenshot 2
InstantPay screenshot 3
InstantPay screenshot 4

InstantPay

Shipped

Fintech payment wallet built as a monorepo with a modular backend architecture — service, repository, and adapter layers with clean separation of business logic.

  • Secure webhook processing with HMAC signature verification and idempotency checks to prevent duplicate transactions.
  • Transaction lifecycle state machine (Processing → Success/Failed) enforcing valid state transitions.
  • Integer-based money representation eliminating floating-point precision issues.
  • CI/CD with GitHub Actions and Docker; route-level code splitting on the frontend.
TypeScriptNode.jsPostgreSQLPrismaTurborepoDocker
GitHubLive

Skills

Languages & Core

TypeScriptJavaScriptGoC++SQLNode.jsExpressHonoReactNext.jsTailwind CSS

AI & Automation

LLM IntegrationPrompt EngineeringAI Workflow DesignContent Generation Pipelines

Infrastructure & DevOps

AWS (EC2/S3)Cloudflare WorkersNginxPostgreSQLMongoDBRedisDockerGitTurborepoGitHub Actions